By now, you have undoubtedly seen press reports claiming that North Korea may have conducted a pair of clandestine nuclear tests in April and May 2010.  The reports are based on a forthcoming paper by a well-known Swedish radiochemist, Lars-Erik De Geer.

I don’t buy it. At least not yet.

Look, I would be the first person to jump at the possibility that the CTBTO’s IMS detected a well-hidden nuclear test. I am one of the few cranks out there who believes the  DPRK may explore boosted fission weapons, which De Geer believes accounts for the pair of alleged tests.  But, as I told Nature’s Geoff Brumfiel, the paper  ”doesn’t feel right to me.” (Science & Global Security has made available an advance copy to me; the issue will be published in March.)

What follows is my best accounting of what I see as some methodological problems with a very interesting, but ultimately unpersuasive paper.

Let’s get a bunch of stuff out of the way first. De Geer is a well-respected Swedish radiochemist with strong ties to the CTBTO.  He’s also a pretty nice guy and has been generous in sharing a bunch of radiochemistry on the Chinese atmospheric nuclear testing program with me.  He’s not a bad sort, even if there are a lot of people in Vienna wondering why he just published this paper without workshopping it a bit at the VIC.

The paper was also peer-reviewed.  Although I believe some of the problems I will outline ought to have been raised in peer review, it seems plausible that one or more peer-reviewers were so focused on the very difficult radiochemistry calculations that they didn’t step back and think about the paper in context.  I don’t know anything about radiochemistry, so it’s easy for me to think about the paper in context.  That’s all I have.

Questionable Methodology

My concerns about the paper are simple to explain.  The paper relies on radionuclide monitoring to detect a nuclear explosion, but the general view among experts has been that radionuclide monitoring is imprecise enough that it should only be used to screen events. So, for example, if there is a seismic event, then the presence of xenon or other fission products might help persuade states to seek a special inspection.  But it doesn’t work the other way around. That is why, for example, the South Korean government cited the lack of seismic activity as a reason to dismiss the xenon measurements when they were initially reported in 2010.

De Geer revisited the 2010 debate and found two interesting sorts of data: xenon measurements at a national radionuclide monitoring site near Geojin (South Korea) and an IMS site near Takasaki (Japan) and barium/lanthanum measurements at CTBTO IMS sites near Usurriysk in Russia and Okinawa in Japan.  (Only lanthanum was detected at Ussuriysk.)  All these measurements occurred between 13-18 May 2010.

De Geer, in general terms, makes two arguments — one relating to analysis of xenon isotope ratios at Geojin and Takasaki, the other relating to the presence of fission products barium/lanthanum at Ussuriysk and Okinawa.

My understanding, based on conversations with radiochemists and a review of the pertinent literature, is that the backgrounds for xenon releases are so bad (and getting worse) that atmospheric mixing essentially eliminates the possibility of using isotopic ratios to
discriminate among xenon sources. Japan and South Korea have large numbers of nuclear reactors.  The background should be quite poor. Even the most encouraging results — studies in 2006 and 2010 that list Martin Kalinowski as the lead author — indicate that it is not possible to discriminate xenon from a explosion against that from a load of fresh fuel that has been exposed for only a few days.  As we will see, there is a plausible scenario for a fresh fuel load at the same time.

I am also uncomfortable with how De Geer approached the task of modeling the xenon ratios.  De Geer clearly modeled a hypothesis of a single test — but the isotopic ratios indicated rejection of his hypothesis.  So he then postulated a second test, placed it in the same chamber to explain the unusual xenon ratio, and adjusted the time between tests to produce the correct xenon cocktail for release.

There is no a priori reason to assume North Korea would conduct a pair of tests separated by a month in the same chamber – previous DPRK tests branched off a main tunnel into separate chambers —  other than that just happens to fit the data.  As one colleague noted, this is rather less like Occam’s Razor than Occam’s Toothbrush.

Finally, De Geer places enormous confidence in atmospheric transport modeling – using weather data to infer the location and source term of the radionuclides.   De Geer was a coauthor on a paper claiming that the CTBTO station in Yellowknife, Canada, had detected xenon from the 2006 DPRK test.  There is some discussion within the technical community about whether it is possible to exclude other sources, including the relatively nearby medical isotope production center at Chalk River. (There are many sources of xenon, including routine reactor operations and the  production of medical isotopes.  Chalk River is a massive producer of medical isotopes and some experts think the xenon detected at Yellowknife might have been from the 2006 DPRK test, Chalk River or some combination of both.)

De Geer’s observation that the station at Okinawa detected the fission product barium is intriguing.  (Lanthanum alone is not — a spike in Germany in 2004 turned out to be from a military contamination exercise.)  Taken together the barium/lanthanum readings at Okinawa and Ussuriysk  do seem to indicate fission.

If the reading at Okinawa is not a false positive, then something interesting happened.  That appears to be one reason why Frank von Hippel, who is quoted skeptically in the Nature article, notes that there must have been some sort of fission explosion.

Modeling Alternative Hypotheses

My colleague Ferenc Dalnoki-Veress and I are currently working to formulate and test a series of these alternate hypotheses.  The most promising candidate so far is Japan’s fast breeder reactor at Monju, which began operations with a fresh load of fuel on May 6.  Shortly thereafter, on Thursday and Friday, there were a number of alarms — reports differ about how many and what type — that seem to indicate problems with the fuel and leaks of radioactive gas.

Japanese authorities reassured the public that these were false alarms, but perhaps they were mistaken.  Monju suffered a serious accident in 1994 that Japanese officials attempted to cover up.  The resulting scandal kept Monju shuttered for fifteen years — until May 6, 2010.  The pressure on certain Japanese officials not to admit further problems must have been immense.  As it was, Japanese officials delayed announcing the false alarms and were issued a verbal reprimand.  What if the alarms weren’t false?

I am not saying this is what happened.  Ferenc and I are going to model this and other scenarios.  Perhaps, at the end of everything, a DPRK test will still be the most likely source.  But the existence of a plausible scenario that would be very difficult to distinguish from a nuclear explosion — a fresh load of unusual fuel exposed for only a few days — that was not examined in De Geer’s paper suggests that perhaps it would have been best to delay publication.

Ferenc and I are going to start churning through a series of questions.  Once the paper is released, you are invited to participate! Our work  is focusing on three questions:

1.  Modeling a series of leaks from Monju that might account for the fission products and the xenon, as well as continuing to develop other plausible hypotheses such as radioisotope production at the DPRK’s IRT-2000 reactor.

2. Attempting to recreate De Geer’s atmospheric transport model with different software and data packages to try and gauge the uncertainty in the modeling.

3. Determining how much freedom De Geer permitted himself by allowing two tests in a single chamber separated by a month.  With tests separated by anywhere from 1 day to 1 year, is there any xenon outcome one couldn’t engineer?

The overall goal is to try to assign some sort of confidence judgement for the hypothesis of a pair of DPRK tests in a single chamber, relative to other explanations.  Nuclear testing may turn out to be the most likely explanation.  But  policy-types should not take this at face value just yet.

Why Didn’t the USG Reach the Same Conclusion?

I should say, in closing, that I am also worried about publication bias.  Shortly after the xenon detection at Geojin, the ROK dismissed the possibility of a North Korean test on the basis of a lack of any seismic data.  The  United States looked into the issue as well and also dismissed a North Korean test, though on what grounds I do not know. Of course, no one publishes negative results and, in this case, there is good reason official inquiries were conducted on a classified basis.  Still, I would like to understand why other competent radiochemists reached a different conclusion than De Geer.  Perhaps De Geer’s work is better, but perhaps it is also simply an artifact of his very carefully engineered scenario and choice of modeling tools.

As a policy analyst, rather than a technical expert, I can’t referee debates about atmospheric transport modeling or the analysis of xenon isotope ratios.  But a policy analyst should be sensitive to areas where technical experts disagree about the confidence of certain tools and models.   We can observe that there are significant uncertainties in the data and tools brought to bear on this problem.  De Geer concludes “The probability … that a low-yield underground nuclear explosion was carried out on 11 May 2010, or possibly, the day before, is significant.”  I think our task now is to ask “Significant compared to what?”

Sorry to poor Josh Pollack for my stealing of his inspired image choice when this controversy first appeared in 2010.

Did you know there is a Scud missile in the lobby of the Defense Intelligence Analysis Center.  A colleague sent along this little video, noting that if you look closely, there it is at 0:13! (Apparently, they also use it as a rally point as in “meet by the Scud”.)

I know that the United States acquires foreign military equipment, including Scuds, on a fairly regular basis.  Still, I’d love to know the story on this one.  There has to be a plaque or something.

What amazing schwag!  This ranks right up there with the B61 in the Forrestal Lobby.

Speaking of which, I tried Googling “B61″ and “Forrestal” just to make sure my memory was correct.  Nothing.  Nada.  Washington has all these buildings with military trophies and such — a Scud at DIAC, a B61 at Forrestal, the centrifuge casing Paula DeSutter tried to take home.  (Where is that, now?)

We really should compile an Arms Control Wonk Guide to these sorts of exhibits in the greater Washington DC area.

Both Toby and I raised a similar question about the possibility of politics in North Korea, which I thought might temper ever so slightly Pollack’s stark but ultimately compelling conclusion.  You can read the reviews, as well as Pollack’s response.

For the purposes of a blog post, I wanted to pick up on a question posed by Pollack to illustrate why politics might matter.

Then I want to share a picture of Kim Jong Un wedged into tank.

In his response to the essays, Pollack notes that “There has been very little commentary in the immediate post–Kim Jongil period on the nuclear weapons program …”

Although Pollack is primarily interested in what drives the DPRK to possess nuclear armaments, I’ve been wondering about the impact of the post-Kim Jong-il period on what’s been called “onward proliferation.”

I wrote  a short op-ed for Kydodo News on this very topic, which also came up at a meeting I attended recently.   And, while Kim Jong-il was still alive and kicking, Joshua Pollack — Jonathan’s son – considered carefully the question of why North Korea sells what it does to whom for the excellent blog, 38 North.

In the United States, politicians openly compete, often using specific policies to distinguish themselves from their opponents. A perfect example is Mitt Romney’s sudden hostility to the use of a mandate to expand access to health care.  If President Obama is for something, chances are that Republicans need to be against it.  It works the other way too, of course.  Contrast is such a good thing that candidates may actually exaggerate their degree of disagreement, despite the fact that Robert Gates could quite comfortably serve both George W. Bush and Barack Obama.

In North Korea, competition is not open and, as far as I can surmise, concerns competing patronage networks more than ideological groups.  This is not an unusual state of affairs.  Political parties in  the United States used to be more ideologically heterogenous, or so I am told,  especially in the machine politics that dominated urban areas.  Even today, the role of interest groups preserves some ideological heterogeneity within political parties and the political beliefs of one’s parents remain an excellent predictor of partisan affiliation.

Proliferation in North Korea, as far as I can tell, is a family affair conducted by one of these patronage networks.  This blog has taken a special interest in one Jon Byong-ho and his son-in-law (maybe) Yun Ho-jin, who Josh Pollack described as the “dynamic duo” of North Korean proliferation.  It was Jon who is alleged to have written that letter to AQ Khan.  And it is Yun who seemed to be in charge of procurement for North Korea and its clients. These two characters, along with a few others, appear responsible for many of North Korea’s more objectionable activities involving ballistic missiles and nuclear weapons.

Jon is also a senior figure in the DPRK regime, among the ten most influential North Koreans, as far as DPRK-watchers can infer from issues of protocol, including pictures of who gets to stand with Kim Jong Un, committee memberships and laundry lists like the  funeral committee for Kim Jong Il.  North Korea’s proliferation activities are more likely structured to support Jon’s machinations in Pyongyang, not the other way around.  Whether Jon sees proliferation as a source of hard currency to buy influence or strategic technology to impress fellow hawks, or both, is difficult to say.

That makes it very difficult to imagine how the  leadership transition could affect proliferation.  At least two sources (one , two) describe Jon as an ideological conservative, especially compared to someone like Chang Song-taek. (By the way, I know my Romanization of Korean names is inconsistent and am open to suggestions about fixing that.)  Chang may emerge as a sort of regent being married, as he is, to Kim Jong-il’s sister.  But Jon, too, is reportedly cl0se to the Kim family and is a part of the so-called “1980 Group” that rose to power with Kim Jong-il.  Whether Chang and Jon will act as allies or enemies is not clear to me — and, in North Korea’s opaque system, it may not be entirely clear to them, either.    There are people who follow such things very closely who certainly will have plausible answers to such questions, but I think it is important to understand that the internal politics of an authoritarian system are almost certainly opaque to the participants, as well as the outside world.  There will be an inevitably jockeying for influence in North Korea, even if only within the context of the preservation of the Kim regime. But many will play their cards close to vest.  We should not expect to do better than the participants in handicapping the horse-race.

North Korea’s approach to proliferation, therefore, may be determined by factors that we simply cannot see and that are, to a first approximation, not “strategic” in any geopolitical sense.   We can not predict them, I suspect, based on a “rational” model that places international interests ahead of domestic ones.

It is perhaps the case that better informed North Korea watchers could do much to answer some of the questions sketched here.  They are certainly invited to do so.

***

Ok, enough of the intellectually heavy stuff.  Let’s look at a pictures of Kim Jong-un wedged into tank.

“Yo, I am stuck.  If you don’t get me out of here, I am going to open a can of juche on your asses.”

There is also, I have been told, a picture of him wedged into a fighter jet — but the images released from his only visit to a KPA Air Force Unit that I could find didn’t deliver.  There is, however, a picture of him looking at a giant fish.

“Yo, what are the rest of you guys gonna eat? Just kidding with you.  It’s all good.”

Sad news today, as Bruno Tertrais has emailed me to say that Thérèse Delpech, doyenne of the French strategic studies community, has passed away.

The last time I saw Thérèse, my wife and I were standing at the TGV station in Avignon, trying to figure out how to purchase a ticket with a North American credit card.  (We ended up stealing the short ride to Aix.)

We found ourselves in Avignon after touring  the decommissioned French fissile material production facilities at Pierrelatte and Marcoule.  I had met Thérèse at several meetings over the years, but that trip to Provence was the first time I really understood how special she was.  I was seated at a lunch with a few French experts, including Thérèse and Bruno Tertrais.  Thérèse could be combative in meetings, so it was with a little trepidation that I sat down.  The previous time I had seen Thérèse was in Paris where she was not very impressed by some of things my traveling companions were saying about nuclear weapons. “Well, it is your deterrent,” I remember her saying, with “your deterrent” sounding exactly like “your funeral.”

Lunch turned out to be pure magic.  Thérèse enjoyed talking about wine and philosophy as much as nuclear weapons.  That lunch is one of my favorite memories, as Thérèse turned her formidable intellect to all the things I really find interesting in life.   (The wine was also excellent!)  By the end of that trip, I was completely charmed. The next thing I knew, we were all in Avignon, with Thérèse troubleshooting the TGV before bidding us farewell.

There is something strange about finding out how much you enjoy someone’s company only to never see that person again.  I noticed that, recently, Thérèse had been traveling less and was cool about committing to a conference I’ve been planning.  It never occurred to me, although it should have, that there might be a reason she was staying close to home.  Thérèse was such a presence that I simply couldn’t imagine one day she would be gone.  It would be like waking up in Paris only to see someone had taken down the Tour Eiffel.

Thérèse was a very special person.  She will be missed.

Josh Pollack has an amazing article in Playboy (of all places) that identifies AQ Khan’s so-called fourth customer: India.

The image of a Pakistani magistrate revoking AQ Khan’s pardon intercut with scenes of Josh cavorting with bunnies at the Playboy Mansion is wonderful to contemplate, even if neither is a likely outcome.

Still, a boy can dream.

Much of the evidence assembled by Josh has been hiding in plain sight, but the notion that India might be the much-debated “fourth customer” so fairly boggles the mind that some people can’t get their heads around it.  In deference to the cognitive dissonance, Josh spends a considerable part of the article dedicated to Khan himself, the actual man behind the myth that he carefully assembled through the exertions of journalists of all sorts. As biography, it is wonderful reading, as Josh meticulously demonstrates that Khan would have no qualms taking India as a client.

The evidence is becomes overwhelming once one admits this strange possibility.

Let’s get one thing straight: The evidence is incontrovertible that India was a customer of the Khan network.  South African court documents state that South African elements of the network sold UF6-resistant flow meters to India.  Moreover, Pakistani officials, including Khan himself, have openly stated that India acquired centrifuge design information from the network, usually blaming deceased individuals within the network for operating independently. We all have known about these relationships for some time, as well as the fact that the Indian centrifuge design bears a family resemblance to Pakistan’s P2.

What Josh has done is make a further claim: that rather than being any customer, India is the customer everyone guesses at — the so-called “fourth customer” after Iran, Libya and North Korea. Usually, one reads that the IAEA hypothesized the existence of a fourth customer on the basis of a series of missing shipments — the IAEA concluded the shipments may have been diverted to a fourth country.  What Josh discovered in his research is that the IAEA had a second, more revealing, reason for believing in a “fourth customer” — Khan and his associates actually used that very term to protect the client’s identity:

 “Members of the Khan network would refer to ‘the fourth customer,’ ” says Heinonen. “It was their code language. We still don’t know who they meant.”

The unusual degree of secrecy is striking: If Khan could admit to sales to Iran, Libya and North Korea, what country could possibly be so sensitive that it must not be mentioned even internally?  When I read this, I immediately thought of the second condition of AQ Khan’s pardon: “the pardon would be ineffective if an evidence of illegal export of “nuclear-related material” to some country other than Iran, North Korea or Libya was found.”

Pakistani officials, including Khan himself, have done everything possible to prevent this information from leaking.  Iran and Libya? Fine. North Korea? No problem.  India?  That’s a different kettle of fish. That would be a very big problem. Khan, even as he was burning customers and patrons alike, as well as leaking documents to accuse other Pakistani officials of taking bribes, protected the identity of fourth customer at all costs.  No wonder.

We know from leaked cables that Pakistani officials have claimed, perhaps not too convincingly, that they would be willing to make Khan available to the IAEA but for that troublesome pardon.  “The facts, said Kidwai, were clear — Khan had admitted his guilt and received a presidential pardon,” according to a leaked State Department cable (Warning! Wikileaks!) documenting a meeting between Khalid Kidwai, Director of the Strategic Plans Division, and the US Ambassador to Pakistan. The revelation that India was the fourth customer may alter that calculation, if not legally than politically. I suspect, now that this possibility is out in the open, we may find further evidence beyond the sale of the flow meters that makes clear Josh is correct about India’s identity as the fourth customer.

So, give the article a read — there is nothing unsafe for work on the story itself, although I wouldn’t go clicking on any links at the office.  When you do so, try  to set aside all your preconditions about the geopolitical rationales for proliferation and, instead, look closely at Khan the man. It isn’t a very pretty picture, but I think you’ll find it very illuminating.

Oh, and the pictures of LiLo aren’t half-bad either.

This is, purportedly, a very interesting Urdu-language article in a Pakistani paper, The News.  I don’t read Urdu, so as far as I know, this is just a recipe for chicken biryani.

According to some translations floating around, the article cites a “Western diplomat” claiming that former Pakistani President Pervez told US officials that Pakistan had developed “among the world’s smartest nuclear tactical devices.”

Now, since nuclear weapons don’t take IQ tests, I think “smart” in this context means “neatness or trimness of appearance.” In other words, miniaturized.

I don’t know if any of this is true, but the Pakistanis seem to be making a lot of noise lately about their tactical nuclear stockpile.  “Look at us!  We have tactical nuclear weapons!”

Mark Hibbs, in a forthcoming article for the Bulletin of Atomic Scientists, mentions that, during his most recent trip to Islamabad, he was directly told that Pakistan was developing very small, low-yield nuclear weapons.

A few weeks after his visit, Pakistan tested a short-range artillery rocket, Nasr. In case you had any doubt, the Pakistan’s official press release stated that the Nasr “carries nuclear warheads of appropriate yield.”

The Nasr is an artillery rocket — I have never before encountered this term Battlefield Range Ballistic Missile or BRBM. Usman Ansari in Defense News quoted a Pakistani academic, Mansoor Ahmed, arguing that “the diameter size of Nasr suggests that the warhead would be less than 1 kilogram, and would be of sub-kiloton range, suitable for battlefield use and could be a fission boosted sub-kiloton fission device.”

I can’t find reliable data on the Nasr, so I used the Mark I Eyeball to observe that the Nasr (top) looks a lot like an M30/31 MLRS rocket (bottom). Both rockets have a similar range.

The M30 carries a 90 kg unitary penetrator and is about 20 centimeters in diameter.  That’s actually too small for even the smallest US nuclear weapon ever designed, the W54. (Aw, look at the little feller.)

Now, of course, maybe the Nasr is 30 centimeters in diameter (I certainly am not/not claiming to be able to eyeball rocket dimensions to within a few centimeters) or Pakistan just made a really little warhead.

I really wonder whether Pakistan could develop such a small warhead with any confidence. The general view has been that Pakistan would probably need testing to develop a miniaturized plutonium implosion device, to say nothing of the sort of boosted little devil we are talking about. I am no warhead designer, but here is how the National Academies described the situation in 2002:

Pakistan similarly could manufacture and stockpile its enriched uranium fission weapons without further testing, and it could make progress toward a plutonium implosion weapon (perhaps even producing and stockpiling one of simple—and inefficient—design, in which it could have some confidence).

Now, as I have noted before “simple—and inefficient—design” is a term of art that usually means too big for a missile, let alone a little rocket like NASR. I am skeptical, I must admit, but still intrigued. This is a smart little puzzle, isn’t it?

It seems fitting that I should end my time as a Washingtonian with a farewell visit to the Big Hunt for a good-bye Happy Hour.  It’s been a very long time, so in case you don’t remember the particulars — the Big Hunt is located at 1345 Connecticut Avenue NW.  We usually show about 6-ish.

If you miss it, don’t worry — I’ll be back now and again for meetings and so forth.  But it will be nice to formally close out my time in Washington.

Well, for one thing, Karber’s claims are utter nonsense. For another, Karber is unbelievably successful at generating unwarranted publicity.

Sure, China has lots of tunnels.   But all of Karber’s sources about fissile material production are based on a mid-1990s Usenet posting by an internet troll.

Actually, it’s worse than that, but this will take some time to explain.

Let’s start with my letter to the Washington Post, which the Post refused to publish. The short version is that China simply doesn’t have enough plutonium for all those warheads, tunnels or not.

 I am appalled by William Wan’s article entitled “Digging into China’s nuclear tunnels.”

Nuclear weapons are not made with tunnel spoil. They are made with fissile material.  China does not have enough fissile material to support a stockpile of the sort that Dr. Karber imagines.

Based on publicly available information about China’s nuclear weapons testing, we know that Chinese nuclear weapon designs since the 1970s make liberal use of plutonium. China operated exactly two nuclear reactors for the production of military plutonium through 1991. Open-source estimates reliably band China’s production of plutonium at 2-5 metric tons.  Classified Department of Energy estimates, leaked to the press, provide a narrower band of 1.7-2.8 metric tons.  (Hui Zhang, a former colleague of mine at Harvard who previously worked in the Chinese nuclear weapons establishment, calculates Chinese production as being on the low end of that estimate in the most recent International Panel on Fissile Materials report.) Using a conservative estimate of 4-8 kilograms of plutonium per warhead, that yields a total force of probably no more than 375 warheads, with an extreme upper bound of no more than 700 warheads.  This figure would be for every plutonium-based nuclear weapon China ever produced, not merely those in the current stockpile.

Declassified US estimates of China’s current nuclear weapons stockpile from the mid-1990s, not surprisingly, place the size of China’s nuclear weapons stockpile between 200-300 nuclear weapons.  In more recent years, Chinese officials have stated that China possesses the smallest nuclear weapons arsenal among the so-called P5 states.  The UK and France are believed to have less than 200 and 300 nuclear weapons, respectively.  Notice how all these numbers are copacetic?

So where does Dr. Karber get his wildly divergent estimates? Nowhere does Mr. Wan mention that Dr. Karber’s “analysis” of China’s plutonium production relies on a few Chinese blog posts that discuss a single, anonymous 1995 Usenet post, subsequently plagiarized by a Singapore University student, that is so wildly incompetent as to invite laughter.  (I have mocked this essay repeatedly on my own blog, Arms Control Wonk.com.)

Actually, Dr. Karber doesn’t mention this either.  His research ended with the Chinese blog posts, which is something that no responsible scholar would do.  A real scholar would have traced the blog posts back to the original Usenet posting, then back to the article in a Hong Kong dissident publication that started this nonsense and then gone to the library (I know, such a chore!) to make a copy.  Dr. Karber did none of these things.  My colleague, Dr. Gregory Kulacki, however, did exactly that.  We will be making the original document available, but let me simply observe that it is very clear that this is simply not a reliable source for the size of China’s nuclear arsenal.

If I take any solace out of this pathetic episode, it is that Dr. Karber’s students will have learned first-hand how not to do research.

Jeffrey Lewis
Monterey Institute, Washington, DC

The Post Letters editor Michael Larrabee objects, by the way, to my claim that the Post “refused” to print the letter — although not so strongly that he would agree to publish it. Larrabee also would not agree to publish a letter from Gregory Kulacki. It will be interesting to see if the Post will permit any criticism of the story to appear on its pages.

I also received a form letter from the reporter, William Wan, similar to one he sent to several other people in the arms-control community.  Wan, by the way, was aware of much of the story I am about to relay, since Gregory Kulacki talked extensively with him in advance. Wan didn’t let that stop him, of course.

Gregory has detailed a lot of what I am going to say in a series of blog posts at All Things Nuclear: Research in the Internet Age: Karber and China’s Nuclear Arsenal, The September 1995 Trend Magazine Article, and Prof. Karber Adjusts His Report on China’s Nuclear Arsenal.

You should totally read Gregory’s posts. (While you are at it, also read commentaries by James Acton and Tong Zhao.) But this was a collaborative effort, and I’ve been looking forward to telling what I think is a pretty interesting yarn that ends in a very embarrassing way for Dr. Karber. And, since Gregory is so nice and professional, I sort of feel someone should tell it my way. You know, mean.

Now, Karber basically omits any discussion of the declassified US intelligence estimates of the size of China’s nuclear weapons (200-300) and plutonium stockpiles (1.7-2.8 metric tons) from the mid-1990s which, when cross-referenced against open-source reconstructions of plutonium and highly enriched uranium production from Albright, Wright and Gronlund, and Zhang, really should end this discussion. There is nothing in Karber’s sources that would undermine these estimates.  Instead, Karber acts like the only publicly available estimates are from old DIA projections or the rantings of NGO-hippies like NRDC or me — ignoring the fact that we both rely heavily on US government estimates that pretty clearly demonstrate there just isn’t enough plutonium to go around. Hans Kristensen has a post using this method, titled simply No, China Does Not Have 3,000 Nuclear Weapons.

On To The Main Event

This is a slide from Karber’s presentation at a dinner seminar hosted by the Nonproliferation Policy Education Center on September 26 titled “China’s Underground Great Wall: Have We Underestimated Beijing’s Strategic Forces?”  As “Chinese statements on nuclear weapons,” Karber cites a few Chinese blog posts, one of which really got my attention.  When Gregory sent it to me, I immediately emailed back — “It’s the return of Yang Zheng!”

Now, my Chinese is lousy but even I recognized Yang Zheng’s essay. For those of you who don’t remember, some so-called experts cite an essay by a Singapore college student, Yang Zheng, as evidence that China has thousands of nuclear weapons.  (I have dealt with this subject repeatedly.) Suffice to say, this essay is completely incompetent — the list of fissile material facilities is just totally wrong. Yang claims this information all comes from a leaked Chinese military document provided to a Hong Kong dissident magazine called The Trend.

Now, as Gregory has already noted, Karber cites a bunch of Chinese-language blog posts that are best described as general interest discussions of the claims made in the Yang Zheng essay, not as definitive statements of Chinese nuclear-weapons holdings. That’s not to say the authors aren’t clever, but the authors are not professional analysts and have no privileged access to information. They are just random people saying things like Hey! I read on the internet that China has 3,000 nuclear weapons. Is that true?

About the same time, Gregory and I received a forwarded email from Karber defending his estimate, in which he wrote that “The 2350 number was originally reported to have come from a leaked PLA document in 1995 …” Now, mind you, Karber didn’t have a copy of that “leaked document” or even the original article in The Trend. He was just repeating something he found on the internet.  But, since he made such a clear reference to the Yang Zheng essay and the 2350 estimate, Gregory and I decided it was time for someone to actually do some research and get a copy of this article.

The first thing I discovered was that Yang Zheng was a bit player in this whole fiasco — he had plagiarized his essay from this 1995 Usenet thread. Do you remember Usenet? Probably not if you had a life in the mid-1990s. Usenet was an internet bulletin board.  As I say, Yang either plagiarized the essay — either that, or  someone cleaned it up and falsely attributed it to him. Otherwise, Yang’s main contribution was to drop into the discussion and start throwing around racist comments, calling some of the other posters “Bananas” — yellow on the outside, white on the inside.  Classy.

The actual author of the incorrect list of fissile material production facilities was an anonymous poster named Ma Tuowen — that’s “Mark Twain.” How clever.  It turns out Ma was posting anonymously because “I don’t want my boss to know I’m wasting his time.”  (Ma’s areas of interest range from Chinese nuclear weapons to strategies for the computer game Civilization.) He had quite a history of saying irritating things.  A very young Dwayne Day, now at the National Academies, got very irritated with Ma and his “poorly-written, bombastic, incoherent bullshit.” (D-Day does not suffer fools. It is one reason he is awesome.)

Then I noticed something really interesting — Ma didn’t actually read the original Hong Kong article himself.  (Yang Zheng, in his plagiarized essay, claims Ma is the source of the leaked document, which is totally wrong.)  Ma was just speculating in response to a question by yet another reader, a Chinese mathematics professor named Li Xaolin.  I emailed Li, who is a nice guy and remembered the discussion.  “Now I remember, it was 1996, when Li Denghui (Tenghui Lee), then the President of ROC announced ‘two countries’, there was a debate in soc.culture.china, soc.culture.taiwan. Some Taiwanese claim that China was no match in nuclear arsenal to the west with only a couple of hundreds of warheads.” Ma just read Li’s summary of the article in The Trend, then made up his bogus list of Chinese fissile material facilities, that later appear under Yang Zheng’s name.

So, it turns out the entire discussion is based on a very casual summary of something a mathematics professor thinks he read in Hong Kong paper. Obviously, we needed the source of all this nonsense.

Here is where it gets interesting.

Gregory headed to Hong Kong to track down the original document and try to make contact with the author.  Gregory went through every issue of The Trend until he found it.  And there it was! At that point, it became clear that this was not a very reliable source of information.  The article claims that China, for example, has six Xia-class ballistic missile submarines and was conducting undetected 15-kiloton nuclear weapons tests in Sichuan.  Yeesh! Gregory has placed a copy online and also made available a copy to Karber.

Most importantly, the article in The Trend does not cite “a leaked PLA document in 1995″ as the source of the numbers.  The mathematics professor just made a mistake.  All these people have simply been repeating his summary on an internet bulletin board, but the numbers in the article come from someplace else.

Then Gregory noticed something even more interesting. The Trend has a slightly more formal sister publication.  It turns out the two publications each carried an article on the topic of China’s nuclear weapons within an issue of the other, a topic neither had covered before or would cover again.  Although the authors use different pseudonyms, the two articles seem to Gregory to be written by the same person.  Someone who, apparently, lives in the United States.

We are still trying to find the original author of the article in The Trend. But it may not matter because we have found the source of his estimates.  It appears that the numbers in The Trend were copied from a 1986 article in the English-language publication Navy International, written by an American named Bradley Hahn, who published extensively on China’s nuclear forces in the early 1980s.  We are still trying to reach Hahn, but it seems that the author of The Trend piece simply took details from Hahn’s article in Navy International 91:10 October 1986, pp 624-630, and tweaked them slightly.  Let’s compare:

* The Trend: “In the CCP nuclear arsenal, 40% are nuclear bombs, 34% are strategic nuclear missiles, 25% are tactical nuclear missiles, 1% are nuclear mines.”*

* Hahn: “An estimated breakdown of how they employ their current nuclear arsenal reveals about 44% bomb, 34% strategic missile, 20% tactical missile and about 2% mine warheads.”

There are lots of other similarities. Put them together and it is a total cut-and-paste job with some salacious details about nuclear weapons testing thrown in.

As far as we can tell, Hahn had some pretty amazing access for an American in the early-1980s. But I think some of his Chinese sources pulled his leg a bit.

Guess how Gregory noticed the overlap between The Trend and Hahn?  He didn’t — one of Karber’s Chinese blogger sources did! Gee, I guess bloggers sometimes know what they are doing. Unfortunately for Karber, this means he and his students knew, or should have known, they were dealing with recycled American estimates. But Karber didn’t share this little detail.

I think it is important to document all this. As Gregory explains over at All Things Nuclear, Karber is already starting to alter his slides — in particular, the one that I have reproduced in this post — in ways that obscure the source of his “estimates” of China’s nuclear weapons stockpile. Gone are any references to 2,350 warheads, replaced with annual production numbers from the same poisoned source.

No, China Does Not Have 3,000 Nuclear Weapons

Wan quotes Karber saying “I don’t have the slightest idea how many nuclear weapons China really has, but neither does anyone else in the arms-control community.” That’s false — a lot of people in the US intelligence community, as well as in the arms control community, have spent a lot of time and energy developing estimates of China’s fissile material production to bound the potential size of China’s nuclear arsenal. To equate Karber’s know-nothingness with these efforts at real scholarship is offensive.

We actually know quite a lot about China’s nuclear weapons, thanks to real scholarly research. You know, the sort of thing that Gregory did — starting with the simple task of going to the library. It is a little tedious, I confess. But it can be fun — like when I trekked to China’s original nuclear weapons design facility near Haiyan and brought home a lot of footage of Chinese nuclear tests and pictures of bomb mockups.

I don’t know how to describe what Karber and his students did, although “threat inflation” comes to mind.  So do a lot of other not very nice terms.  Gregory chose “incompetent and lazy,” which William Wan used to make Gregory’s legitimate criticisms of Karber’s methods seem overheated.

I like “goat rodeo,” but that’s just me.

Karber dismisses criticisms as though his speculations are just part of a healthy debate. But preparing for a debate takes time and money — it is much easier for Karber to sit in his house in Great Falls and make things up than it is for Gregory to get on an airplane and go to the library in Hong Kong. A participant in an academic debate has a responsibility to do his homework — an obligation captured in the notion of scholarship that Karber shirks by simply implying that real estimates are more or less the same as whatever he happens to find on the internet that day.  By Karber’s reasoning, he could just as easily argue that all those Chinese tunnels house aliens. (Hey, maybe this is where the Chinese store all our Treasury securities!)

So, no China does not have 3,000 warheads. But thanks for wasting everyone’s time.

Sometimes, when that happens, it helps to simply ask questions.  Robert Schmucker and Markus Schiller have been up to the same thing, and send along their list of preliminary thoughts and remaining questions:

On November 12, 2011, an explosion rocked an Iranian military facility just southwest of Tehran. Up to 20 military personnel are reported to have died in the blast, including a General who was regarded as a leading figure in the Iranian missile program.

A few days ago, Paul Brannan of ISIS published a short piece on the incident that also included satellite imagery of the site 10 days after the blast. Michael Elleman of IISS pointed us to this report and asked some very good questions, triggering our interest in the event.

Here are some of our findings so far. They are, of course, only preliminary and might still change with new information in the future.

–    Compared to other solid-propulsion related facilities all over the world, this site is small.
–    The satellite image was taken 10 days after the incident. The extent of any cleanup efforts during these 10 days is unknown, but it seems that much of the debris was already moved.
–    The damage that is visible on the satellite image is massive. Even buildings 150 m away from the supposed center of the blast show severe damage.
–    The damage pattern of an exploding solid rocket motor is known (for example the Pershing accident at Waldheide, Germany, 1985).
–    This damage pattern looks more like a detonation than a solid propellant burn-off or explosion.
–    The ground is charred at 3 different places.
–    One of these places is right in the middle of a former building (in the very southwest of the image).
–    The other two charred marks are located outside of former buildings, but close to them (slightly north of the first mark, and again north of that, on the northeastern corner of the former blue-roofed building in the northwest).
–    The charred marks are large, at least 15 m in diameter.
–    Composite propellants generally do not detonate – they are Class 1.3 explosives. Pure ammonium perchlorate might detonate under certain conditions.
–    With a detonation of this size, one should expect crater(s), especially if the detonation(s) took place right on the ground (as with barrels filled with ammonium perchlorate, for example).
–    No craters are visible on the available satellite image (only some debris that creates the impression of one or two craters).
–    The roofs of high buildings in the distance are heavily damaged.
–    Tall trees that were visible on the earlier image were affected, small trees and bushes seemingly not.
–    When an important leader visits a facility like this, it is standard procedure not to do any dangerous activities that may potentially harm the visitor.

The question now is: What really happened at this site on November 12?

Let me only add a list of possible reference incidents relating to solid-propellant rockets — the 1985 Pershing accident at Waldheide (US accounts use Heilbronn as a place name), the 1988 PEPCON disaster, and the 2003 VLS explosion in Brazil.

Some people are bitching and moaning about the Ploughshares estimate of  $700 billion in spending “on nuclear weapons and related programs during the next ten years.” Many of them are only upset because they are losing the debate over US nuclear weapons policy. In particular, some of the same people screaming about $700 billion are the same people suggesting China might have 3,000 nuclear weapons.  We’re aren’t exactly arguing with Socrates, here.

With that very clear starting point, I am going to try to sort through some of the numbers just so you don’t have to.  Let’s start by recognizing that all of these estimates are simply good-faith guesses.  No one knows, or ever has known, what nuclear weapons really cost or can reliably predict savings from future cuts.  (Well, maybe Amy Woolf can.)  If you are impugning the integrity of a particular participant or handing out Pinocchios, it just shows you don’t know what you are talking about.

Here we go.

The Administration Estimate

The Administration actually provides two apples and oranges estimates of spending on what we might call nuclear weapons and so forth – keeping in mind that “so forth” is an interesting category.

1. Major Force Program 1 Strategic Forces and Atomic Energy Defense Activities

Historically, “nuclear weapons spending” has been expressed in two very imperfect categories in the Green Book: Major Force Program 1 Strategic Forces (sometimes written Strategic Programs) and Atomic Energy Defense Activities (which reports certain spending in the Department of Energy.)

These numbers suck, though not for lack of effort.

MFP1 includes some non-nuclear programs like missile defenses, while it excludes an enormous amount of other relevant spending.  Command-and-control investments are accounted for under a separate MFP, which irritates me.  Command-and-control spending is probably the most important investment we can make in strengthening deterrence.

The Secretary of Defense Task Force on Nuclear Weapons recommended addressing some of the problems with MFP 1:

Funding for strategic capabilities has traditionally been addressed in the account for strategic forces: MFP-1, ‘Strategic Programs.’ Currently this budget program includes both nuclear and nonnuclear program elements. Nonnuclear programs have seen increases in funding as nuclear forces have been decreased in funding over the last 15 years. Some nuclear deterrence capabilities are categorized in MFPs other than Strategic Programs. To avoid further erosion of resources to the nuclear mission, the Task Force recommends that ASD(D) be responsible for funding execution oversight of nuclear capabilities. This is to be accomplished by the creation of a new capability portfolio composed of all program elements (whether currently categorized in MFP-1 or elsewhere in the defense program and budget structure) directly related to nuclear deterrence.

Atomic Energy Defense Activities is more straightforward in that it includes a bunch of cleanup costs. I happen to think clean-up costs should be included in cost estimates, but that isn’t how DOE budgets, so I can see a case for excluding it in this discussion.  (Although it is deeply immoral to continue to pass costs on to future generations.)

If you just take these two numbers, the total easily averages out north of $30 billion a year over ten years.  This number includes some things you would think do not belong, while excluding others.

2. The Administration’s 1251 Report, as amdended.

The 2010 1251 Report as amended, submitted as part of the ratification process for the New START treaty, offers a slightly different estimate, which I suspect is a good-faith effort to replace MFP1 and Atomic Energy Defense Activities with a more realistic number.

The 1251 Report outlines about $125 billion over ten years for delivery systems and another $90 billion or so for NNSA-related expenditures. As far as I can tell, the Administration intended the $200 billion 1251 estimate to be a comprehensive number, not merely new money.  (Representative Turner quotes from the classified 1251 Report, which is a nice trick. The Administration really should declassify the actual 1251 number so we can at least reconstruct its estimate.)

Still, the 1251 estimate  does not include funds for modernizing the Minuteman force, new bombers or new cruise missiles. I can’t blame the Administration for omitting these costs, since no one really knows the cost of these programs.

The 1251 Estimate also includes only a fraction of the monies in “Atomic Energy Defense  Activities.” It excludes funding for Naval Reactors, Office of the Administrator in NNSA, as well as “Environmental and Other Defense Activities,” all of which one could argue about.

The 1251 Report works out to about $20 billion/year.  As far as I can tell, this number doesn’t include anything really profoundly odd, but almost certainly does so through an abundance of caution.

For example, in 1998, the Congressional Budget office estimated the annual cost of a START-1 force of 6,000 accountable warheads at $22 billion.  That’s about $31 billion today. (I was too lazy to try to use differing Defense and Energy deflators.  So sue me.)  CBO’s notional 3,500 warhead force, which is rather closer to what we have today, produced negligible savings (less than a $1 billion/year) since reductions were offset by various modernizations of one sort or another.  (A 2002 study omits a yearly estimate, but suggests the reductions to be undertaken due to the Moscow Treaty were also unlikely to save significant money.) It is very difficult for me to understand how the planned nuclear budget can be 2/3 the 1998 budget in real dollars in light of what CBO found, unless of course the 1251 Estimate excludes some important items.

My major complaint with MFP1 and the 1251 estimate is, as far as I can tell, neither fully accounts for command-and-control spending, which is an important (and expensive) investment.  CBO, on the other hand, did account for some command-and-control spending, which added something like $6 billion a year to the total. Perhaps we would spend this money in any event, but a major recommendation of the 2010 Nuclear Posture Review was to increase investments in the command and control system.  We are, I suspect, spending more than the Administration’s estimate of $20 billion a year by any reasonable metric.

Let’s close our discussion of the “official” number by noting that a ten-year time-frame is completely arbitrary.  Everyone chooses ten years because that’s how many fingers are standard-issue.  Ten years, however, is far too long for accurate estimates, but too short to fully capture the large expenditures on Triad replacement that will be made in the next few years.  SSBN(X) procurement was set to run from 2019-2033 last I checked. Ten years is a handy time-horizon, but I am not sure that it is the relevant one.

The Ploughshares Estimate

Let me start by noting that I am not currently a grantee of Ploughshares. I have been in the past, and some of my colleagues at CNS are, of course. Still, Joe is a close friend. He introduced me to my wife and is my gym buddy. So, take that for what it’s worth.

Ploughshares did three things that seem to bother some people.

1. One Man’s “Related Programs” is Another Man’s … oh, you get the idea

Ploughshares included missile defense and other activities like environmental cleanup for nuclear activities.  One can argue about whether these things ought to be included — some of them certainly are included in MFP1 (missile defense) and Atomic Energy Defense Activities (environmental clean-up).  GAO used to argue with the Defense Department over whether to create a “virtual” Major Force Program for the “New Triad” that would include missile defenses and conventional strike, as well as command-and-control. That was before the budget crisis when everyone started pleading poverty.  I am not particularly bothered by a “nuclear weapons and so forth” approach to accounting as long as one is explicit about the “so forth.”  That’s about $270 billion of the $700 billion figure. Let’s set that aside, since reasonable people can disagree, as long they are consistent and transparent about including those sorts of costs.

2. New Money?

Since no one is disputing the numbers for weapons activities– that means the dispute is whether the number is $125 million/ten years or $385 billion/ten years for delivery vehicles — submarines, bombers, missiles and the like.

Ploughshares noted that it was unclear whether the $125 billion number was new money or not.  This is where I really have trouble with some of the attacks on the integrity of Ploughshares — the working paper clearly notes this ambiguity:

It is unclear how much DoD’s proposed budgets would be new money above base budgets. If the full $125 billion were added, the total estimated ten-year cost of nuclear weapons and related programs could reach approximately $740 billion.

Now, I happen to think the Administration intends the $125 million over 10 years as all money, not just new money.  But that was not clear from the earliest Administration statements.  Moreover, given past CBO estimates in the $20 billion range (or more than $30 billion in current dollars), it was not unreasonable for Ploughshares to guess (probably incorrectly) that $12.5 billion a year was new money.

3. Overhead and Support Costs

Finally, Ploughshares — following Kosiak, Schwartz and others — tried to allocate operations and support costs for US nuclear forces by adjusting MFP1.   As far I as I can tell, both MFP1 and the 1251 estimate exclude very significant overhead costs associated with nuclear forces.  But I also happen to think it is virtually impossible to figure out what the real cost is. That doesn’t make it right to simply pretend these costs don’t exist, so I understand what they were trying to do. Let’s take Minot Air Force Base — if the US did not have nuclear weapons, the Air Force would seek to close Minot by its own admission.  But how much do we really spend on Minot? How much should we charge against the nuclear weapons “budget’? Would BRAC’ing Minot save any money? There is no good way to answer these questions, which I think is why Stephen Schwartz and others are always very modest about their efforts at estimation

This is a good point to make a clarification: What something “costs” is simply not the same as what one proposes to “spend” in a budgetary context, where many costs may be off-budget. And neither is the same as what one might “save” — since achieving savings often entails up-front expenditures.  Dismantling bombs and closing bases is not free. I happen to be of the “hunt for small potatoes” school when it comes to nuclear reductions, unless the Administration makes some very brave cuts to force structure. That’s a Sir Humphrey joke, by the way. Sequestration, of course, may alter this political calculus. But to a first approximation, we might be spending $200, or $500, or $700 billion over ten years, but that doesn’t mean we can save that much.

The annual Ploughshares number of $40 billion a year is definitely higher than other estimates, but the CBO estimate of $30 billion in today’s money for our current force sits comfortably between the Administration at $20 billion and Ploughshares at $40 billion.  If we really are spending new money or there a couple of Nunn-McCurdy breaches on procurement efforts?  It’s a very interesting question whether we might get to $40 billion a year. I think it’s in the ballpark, to be honest. I’d really like to see CBO update the 1998 and 2002 reports to estimate what the Administration could save in this era of sequestration. My own thought experiment was not as disheartening as I had expected.

In Conclusion

In the meantime, if I am talking to other policy wonks, I tend use the Administration’s $200 billion number — but only because this is a tedious f’ing discussion.  I can’t be bothered to explain all the items that $200 billion excludes.  Other than command-and-control, which is my hobby horse. Maybe I’d make a little joke about how $200 billion is the estimate before correcting for the $600 toilet seats or something glib. This entire debate, from a policy perspective, is ultimately irrelevant: the coming cuts will occur to specific program elements, not the general-interest ballpark estimate. The $200 billion estimate isn’t all the spending on nuclear weapons, and certainly not their cost, but it is where the budget-cutters will turn first.

Just let’s not pretend that the Administration’s $200 billion number is anything other than an good-faith, rough approximation that undercounts the full cost of nuclear weapons.  We may find that Ploughshares ends up being closer to the mark. Whichever number you prefer, my advice is not to be an a-hole about it.  There is no reason to send partisan nastygrams or hand out Pinocchios to people who argue that business-as-usual is going to cost a fortune.

Because it will.

Update | 6 December 2011 I fixed some spelling mistakes and edited this sentence for clarity: “Moreover, given past CBO estimates in the $20 billion range (or more than $30 billion in current dollars), it was not unreasonable for Ploughshares to guess (probably incorrectly) that $12.5 billion a year was new money.”  It was originally phrased in passive voice with an odd verb tense that might not even be English.