ArmsControlWonk.com

click on the image for a larger version

Timeline of the 5 April 2009 U’nha-2 Launch

Ok, so I don’t have the amount of confidence in my prediction as the title seems to indicate. But everybody else is making blanket statements so why can’t I?

After returning from the Mid-East on Thursday, I had to catch up quickly on what was going on with North Korea’s missile/rocket program. My post wondering where the launch notification was represents my first attempt to catch up on the DPRK’s activities, after thinking exclusively about Iran for the last two weeks. I’m really uncertain why July 4th was picked up by so many people as the expected launch date. It seems just too soon (I’m talking here about the practical arrangements, not the time necessary to diagnose what went wrong with the last launch and correct it, which is also too short!) and, as far as I can tell, there was absolutely no sighting of the rocket on either launch pad.

Taking a look at the timeline, as determined by various milestones of the previous U’nha-2 launch as reported in the media, indicates it takes about two months to move the rocket to the launch site, assemble it horizontally in an assembly building, erect it on the launch pad, fuel it and wait for good weather to launch. That includes about three weeks warning time North Korea gave by filing a launch notification with international organizations. If we (naively) assume the same amount of time, then they should launch in early August. That assumes the June 1st initial indication of launch preparations (shipping large cylindrical objects by train) that I found was the first one published. It also assumes that intelligence agencies and media organizations are just as efficient as publishing the information as they were the last time. Furthermore, it assumes that North Korea does not need to make any long term preparations that might have been performed for the last launch before the preparations were spotted by the West.

If all of this is true, the first indication should be DPRK’s filing a notification of stay clear zones, probably in about a week. If that happens, everything will be on target for a launch during the first week in August.

Of course, given all these assumptions, I could be proved wrong any day now!

Comment [2]

North Korea is expected by just about everyone to launch at least one (and possibly two?) large missiles/rockets this weekend. And yet, they do not seem to have filed a warning to mariners and airmen notification with the International Civil Aviation Organization (ICAO). This is something they seemed to have been very proud to do over three weeks before the April 5th launch. What’s going on? Have they abandoned their adherence to the Outer Space Treaty already? Or are they just not as far along in their preparations for these (this) launch(es) as they are being given credit for?

Comment [4]

At left, is an Iranian numerically controlled lathe milling machine forming cooling channels in a rocket engine.

It is impossible to tell the scale of this engine — and therefore impossible to uniquely link it to the Safir second stage.

However, machining these channels, as opposed to using a corrugated insert, is a major technological change from SCUD technology.

***

David Wright and Ted Postol have done a really first rate job of analyzing the U’nha-2 and Safir development programs, as exemplified by their excellent article on the U’nha-2. But I think it is important to at least consider an alternative: these missiles represent a much larger portion of indigenous production than just assembling components. This is not to say that Wright and Postol are wrong in their conclusions, only to consider the question.

An International Missile Development Consortium?

North Korea is widely viewed as not testing their missiles enough before they sell them to “client” states. The Nodong missile, which forms the basis for the Shahab-3 and its variants, was tested successfully just once before “being sold to Iran and Pakistan.” This is an unreasonable flight test program and has led many to conjecture that North Korea is either buying complete missiles from Russia, missiles already engineered and developed, or missile components. That could, of course, be very possible and has unfortunate implications for the West’s relationship with Russia. Another alternative of this basic idea, just a small variant really, is that North Korea bought the production line for an obsolete or canceled missile system and modified it to fit its own special circumstances. With this head start, it then formed an international “consortium” with Iran and possibly Pakistan to continue the development. Moving its development program into other countries would have significant advantages for North Korea. For one, while Iranian missile launches are controversial, they do not appear as controversial as the DPRK’s missile tests. This is even more true for Pakistan where any controversy is mainly a regional one.

Such a development consortium would not be the first one ever created. The one I am most familiar with is the Badr 2000/Condor II development program where Iraq, in essence, funded the development of the missile by several other nations. Iraq received a number of contributory production plants that increased their capabilities considerably while failing to produce the desired missile. If North Korea bought the equipment for an obsolete or canceled production line, this would undoubtedly violate many of the rules of the MTCR but might not be as suspicious a violation, especially in a country suffering from the economic catastrophe that was Russia in the 1990s, as selling missile components. After all, most of the equipment could be considered dual use and could appear in separate manifests etc. All the subterfuges proliferation profiteers have used in the past. Importantly, it is much, much easier to reverse engineer a production line than it is to reverse engineer a missile component. After all, once you know the production line components, it is quite easy to buy similar or even exactly the same production equipment else where. The difficulty in reverse engineering is to infer the production scheme.

Too Advanced for Purchasing Production Lines?

An Iranian welding the Shahab engine injection head.
This illustrates the shop-floor know-how that is so
important — and so hard to acquire.

If this happened for the Nodong missile, is it possible it could also happen for an SS-N-6? In fact, it seems even more likely to me that it would happen for these more advanced missile components. The world is full of SA-2 engines, as Iraq showed by purchasing these engines in late 2002 ( see UNMOVIC’s Compendium, volume IV, p. 581.) The closer they get to strategic weapons, the more they come under the control of various treaties. (I’m not sure if SS-N-6 missiles ever came under any of the START etc. treaties, do any of you wonk-readers know?) It is possible that makes them harder to illicitly dispose of. It makes their production lines, however, that much more valuable.

( See my posting on estimating the costs of just the know-how associated with the Badr-2000. That alone was worth $75 million.)

What Proliferators Want

Proliferators, just as would-be producers of civilian products, want access to the technology and they are very seldom satisfied with just components. In fact, the financial inducements needed to entice developing countries into foregoing civilian technology transfer have to be considerable. It seems unreasonable that both Iran and North Korea would voluntarily put themselves into the situation David and Ted suggest, that they only have a finite number of components, and make themselves susceptible to the types of international restrictions that would eventually shut off their missile programs.

Comment [25]

Iran is obviously investing a lot of its political, intellectual, and financial resources into the Natanz enrichment center. But can we come up with a figure for how much it has cost Iran? Perhaps we can estimate it based on costs associated with Western equivalents. I have attempted to do this below. The result is simply a ballpark figure and I should warning you that making budget estimates, just as making sausages, is not a pretty sight!

Warning: What follows might not be suitable reading for the infirm or small children!

Centrifuge Production Know-how…$75 million
Industrial know-how, the techniques actually used by the shop-floor workers, is vitally important for the successful production of any sophisticated item. Unfortunately, it is determined by how much the market will bear. How then should we estimate it? I decided to look at how much Iraq paid (or, rather, was willing to pay) for the know-how to build an advanced solid-propellant missile, the Badr-2000 (aka the Condor II). This know-how cost is explicitly stipulated in the contract Iraq signed with its supplier state: $75 M, after correcting for inflation. It could be argued that Iran might be willing to pay considerably more for the know-how for centrifuge production but any such guess would be just that. (This, as I warned you, is the ugly part.)

Construction of the underground facility…$55 million
ISIS has done a great job in following the construction of the Natanz facility using satellite reconnaissance. Assuming that the holes dug for the “cut-and-cover” enrichment halls are 25 meters deep, then the excavation costs (at $3 per cubic yard) is $7 M. The concrete, at $70/cubic yard, (and assuming floors, ceiling, and walls are 2 meters thick), is then $37 M. Those do not add up to the $55 M but if you assume a 50% “penalty” for working in a desert, then that’s what you get. (Again, ugly.)

Centrifuge production…$140 million
Given that Iran bought the know-how and initial production lines (production equipment not included, ugly!), I am only estimating the cost per centrifuge here. That comes from the cost per centrifuge for URENCO centrifuges as being leased to France. (Ugly, ugly! Let me be clear before somebody takes offense: when I say ugly, I mean my method of estimating is ugly.) You could argue that URENCO centrifuges are more sophisticated and therefore should cost more. Or you could argue that cost is determined by the relative level of sophistication of the production line compared to the past experience of the producer. (That’s what I assume; ugly, ugly, ugly!) I then get a per centrifuge cost of $20,000. Seven thousand of them therefore means a total of $140 M. The one thing that does not make sense is to cost them per SWU; manufacturers produce centrifuges not SWUs.

Grand total cost = $270 million and counting

Comment [5]

I’ve been reading “Caging the Dragon” on the flight to London and it reminded me of a little song from Sesame Street that goes:

One of these things is not like the others

One of these things is just not the same.

What reminded me of that? These three yields:

- Fission yield

- Fusion yield

- Hydrodynamic yield

Can you spot the one that’s not like the others? (Hint: Perhaps it’s easier to try to figure out why two of these are the same.) No fair googling the answer! I didn’t have access to google on the airplane and you shouldn’t use it either.

Spoiler Alert: as I might have expected, keen wonk-readers got the answer right away, so only read the comments if you want to see what others have said.

Comment [9]

Click on the image for a larger version

In a typical one hour period, five US low earth orbit satellites pass over the Port of Rajin

Yesterday, I pointed out that the only way for the West to tell the difference between a cargo container full of uranium gas centrifuges and a container full of pig iron as it was being loaded might be increased security associated with the centrifuges. That, of course, assumes two things: 1) that the West can see the loading and 2) North Korea decides that it wants to include the security. After all, Syria apparently decided to do without anti-aircraft batteries and other security measures for its Box on the Euphrates in order to avoid signaling the US that it should pay special attention to that building. We all know how well that worked: it seems to have fooled the US but Israel, with its spies in Syria (the ones who smuggled out the photos of the construction, presumably from a central office in Damascus) managed to obtain the evidence they needed. Unfortunately, it is doubtful that Israel has spies in North Korea (I guess I could be wrong about that) so we are probably left with using technical means to watch the loading and unloading of ships.

One possibility is to use satellites and another is using UAVs. Let’s consider using satellites, mainly because this post is already getting too long. As the image above shows, the US has a great many low earth orbit (LEO) satellites. The orbital elements for these satellites comes from an amateur satellite observers website and could very well include satellites that have ceased to function. (That’s my fault and not the amateur observers. On the hand, North Korea might not know any better than me which satellites were functioning and which are not.) They do not include the Navy’s NOSS satellites which are used to track ship locations. Those satellites were going to be considered tomorrow when I thought we would look at tracking ships on the high seas as a way of determining if their cargo might be suspicious. It turns out, however, that there might be other, more convenient, ways of doing that (See Allen Thomson’s posting and the responses to that) so I’m going to skip that post.

If you simply require that a satellite be above the local horizon to spy on North Korea, then these satellites supply considerable coverage. ( Click here to see a graph of the satellite elevations over the Port of Rajin during a three day period.) Of course, viewing a scene at a grazing angle—unless you are something like an electronics intelligence satellite which could conceivably pick up the radio chatter between security elements—can prove difficult to interpret. Of course, it’s still possible to see things at angles very close to grazing, just things like buildings and trees get in the way. That and the fact that you are looking at very large distances, perhaps as much 3000 km, so the resolution will be very poor.

With that in mind, I plotted the time between satellite passes where a pass is counted as starting when it appears above a certain elevation and ends when it dips below that elevation. ( This plot is shown here for three threshold elevations: 0 degrees, 20 degrees, and 45 degrees.) North Korea, too, could generate just such a plot and know that critical signatures, such as deploying security forces, would have to be timed to fit in between passes. If satellites can detect signatures at grazing angles, then typical separations between US LEO satellite passes is about 10 minutes. That seems very short, though I have no experience deploying security forces. Its worth pointing out that if the satellite is a photoreconnaissance satellite that normally has a resolution of 10 cm at 200 km, it has a resolution of 1.5 m at 3000 km. There might also be a big decrease in sensitivity of electronics intelligence gathering depending on how they normally function.

Going to a 20 degree threshold substantially reduces the slant range but also substantially increases the time between effective passes. Thus, there is an average gap of 34 minutes between satellite passes in that configuration. Much more can happen in 34 minutes than in 10 minutes: more troops could be better deployed etc.

There are still more practical questions to be answered about achieving “reasonable grounds” for interdicting WMD. Clearly human intelligence would be the best. But that has important problems too, such as protecting the life of the informant if you ever have to prove you had “reasonable grounds.” It seems possible to use technical means to increase your confidence by seeing an unusual amount of security. But the most likely clue might simply be the destination: is it another “rogue” nation? Of course, if all the “law abiding” nations won’t trade with a rogue, perhaps their only trade partners are other rogues.

Comment [9]

The Ports of North Korea

Yesterday, we started to get some ballpark numbers relevant to interdicting WMD shipments. Today, we are going to continue to try to get a feel for the problem of determining whether or not there is proscribed material onboard a ship. The first step, I think, is to considering the volume of shipping going into and out of North Korean ports. The larger the volume of shipping, the harder it will be to notice unusual activities. On the other hand, if there are few ocean-going ships, it will be much easier to keep an eye on what is happening. Tomorrow, we will (finally) get down to the question of actually observing the loading of ships. To jump ahead a little, it is going to prove very difficult to spot WMD related containers and the “reasonable grounds” could very well come from knowing where the ship is going. That will be the final entry in this series.

We (i.e. I) only have a snapshot of that shipping as presented by GoogleEarth though it does manage to convey an amazingly vivid sense of the North’s economy. ( If you need further confirmation of it, check out this image of the Korean peninsula taken at night. )

The GoogleEarth image above shows all the port facilities (and then some) found by North Korea Economy Watch . I have only seen evidence of large, ocean going cargo vessels at a couple of these: Rajin and Yuktae-dong (which NKEW labels Ship Construction/Repair),. Chongjin Shipyard and the other “ports” appear to be mainly coastal vessels with the largest ship I’ve seen measuring less than 30 meters long by 4.5 meters wide. If we assumed these could be ocean going vessels, then each one might hold four cargo contains or about 380 centrifuges as compared to the 61 cargo containers needed for a “standard” cascade interconnect bomb plant. It, of course, makes a lot of sense for North Korea to be emphasizing coastal vessels considering how mountainous the terrain is. Only Rajin appears (at least to on my search) to be the only one with port-based cranes for moving cargo. Yuktae-dong has large quays but the large ship visible has ship-based crane. That, of course, means that the other ports could be used by those ocean-going ships with their own cranes if the harbor is accessible to them. On the plus side, however, the North is missing a lot of the lovely crinkly bits, as Slartibartfast might say, that South Korea has and that might make wonderful temporary harbors.

Now, compare the Port of Rajin with the Port of Busan in South Korea. Busan has a ton of ships just waiting at anchor to get their turn at the docks to load/unload. I’ve circled just some of the 200+ meter long ships at anchor. I stopped when I got tired of doing it. This snapshot might be an even better way of judging the international trade of the two countries than the tonnage of flagged ships since the North and South have similar numbers of ships carrying their flags. The difference in harbor occupation is obviously made up of ships flagged by different countries. One thing the West has going for it in terms of following WMD shipping leaving or coming into the DPRK is, therefore, the dearth of ocean-going shipping. If an ocean going ship is detected, it might be easy to follow.

Of course, it might be very hard to discriminate the loading of WMD related equipment from, say, a load of pig iron. Perhaps the only clue might be the extra security associated with transporting centrifuges or other sensitive equipment. Depending on how efficiently the cranes are run, it might take up to five minutes for each container to loaded onboard though it could take considerably shorter. Of course, if they are loading an entire bomb production plant, this could correspond to up to five hours. All of which could be done at night with minimal lighting on the security force.

Tomorrow, we will consider the frequency of visits over the ports by photoreconnaissance satellites as well as viewing them from UAVs.

Comment [6]

UNSC Resolution 1874:

12. Calls upon all Member States to inspect vessels, with the consent of the flag State, on the high seas, if they have information that provides reasonable grounds to believe that the cargo of such vessels contains items the supply, sale, transfer, or export of which is prohibited by paragraph 8 (a), 8 (b), or 8 ( c ) of resolution 1718 (2006) or by paragraph 9 or 10 of this resolution, for the purpose of ensuring strict implementation of those provisions;

With the passage of UNSC resolution 1874 last Saturday, a young techno-wonk’s thoughts immediately turn to the question of how do you have reasonable grounds to stop a ship on the high seas. (I hope to look at the legal questions in a future post, after I finish with these practical details.) We will only consider some of the issues related to nuclear proliferation here. Some rough numbers immediately jump to mind. These include things like: how many centrifuges can you ship in a single cargo container? How many centrifuges fit on a single ship? What is the “background” to detecting such a ship? (Background in the physicist’s sense: is there a lot of shipping going into and leaving North Korea that could “mask” the WMD ship?) Could you detect the loading of a ship using photoreconnaissance satellites? UAV’s? How could you track it? So here goes:

1) How many centrifuges fit in a cargo container? Libya’s centrifuges were packaged in boxes roughly 2m by 0.6 m by 0.3 m. (All these numbers are rough. We just don’t need better accuracy, as you will see.) We know from images released by the US government that two outer casings were shipped in each box. Since you cannot ship the rotors already in the centrifuge (it damages the ball bearing the rotor rests on; P-1 designs are assembled on site), there is probably another box with two rotors in it that was not shown for nonproliferation reasons. If we assume roughly the same size of box, that means that a standard shipping container (which is roughly 14.6 m by 2.5 m by 2.5 m) could carry, on average, 96 complete centrifuges.

2) How many centrifuges onboard a cargo ship? If you look at the port of Rajin, there is a bulk cargo ship at the only Port in North Korea that I have seen (see North Korea Economy Watch ) with port-based cranes whose below-decks storage area measures roughly 112 × 25 × 21 meters so it could fit 630 such cargo containers below decks. That means a grand total of 60,480 centrifuges could be carried onboard a single ship. Since, as Alex Glaser has shown in his wonderful paper Characteristics of the Gas Centrifuge for Uranium Enrichment and their Relevance for Nuclear Weapon Proliferation, Libya’s design for a “cascade interconnect” bomb factory needed only 5800 centrifuges to enrich over three bombs worth of weapons grade uranium from natural in a year. That would fit in about 61 cargo containers. Clearly, a single ship can pose a significant proliferation risk!

Tomorrow, we will consider some of these other questions.

Update (6 pm): The New York Times reports that the US will not board by force any ships sailing out of North Korea. That was clear from the wording of the resolution: “with the consent of the flag State.” I wonder how many, if any, ocean-going carrier vessels North Korea has. Does anyone know of a publicly available list of ship flagging? My guess would be nearly zero ships owned by the DPRK.

Comment [24]

North Korea is rightly considered an international pariah that not only starves its population while pampering its dictator but also thumbs its nose at the international community’s demand that it not test nuclear bombs or continue its missile development. It is therefore, perhaps, strange to think why the DPRK tested its two nuclear bombs underground.

After all, it is much simpler and less expensive to have an atmospheric test. It also provides a cheap and accurate way of calibrating the bomb by photographing the fireball in the same frame as the sun. China used this method in at least one of its 22 atmospheric tests after the US and the Soviet Union signed the Limited Test Ban Treaty. (That treaty went into force on October 10, 1963 while China’s first nuclear test, an atmospheric one, was a year later.)

Was the North trying to maintain secrecy about the design of its bombs by containing the radioactive particles that might have been used for nuclear forensics? If that was the reason, it seems unlikely that it could successfully hide all the produced radionucleotides, including the noble gases. Of course, the West has, at least to my knowledge, remained very silent on what they have detected. Or was Kim Jun-il concerned about fallout landing on his own people? If that was the concern, they could have positioned it very close to the Eastern coast and waited for a day with a constant wind blowing out of the West.

Or was North Korea, contrary to all international expectations, succumbing to the international norm of only testing nuclear bombs underground?

Comment [19]

Never take anything for granted. I’m afraid I put off something that Peter Brown, who is always on top of the interesting stories in Asia as he writes for the Asia Times, has been bugging me about. Sorry Peter! I thought I knew what the answer would be and I put off calculating the elevation of the U’nha-2 as a function of time far too long as I worked on other stuff. I assumed that the U’nha-2 probably had its second staging event below North Korea’s horizon. It turns out that I was wrong. The rocket was almost certainly 10 degrees, or perhaps even more, above the launch pad’s horizon when its third stage would have burnt out. In fact, the rather simple model I have for the trajectory (which does fit the observed contrail and have its first and second stages splash down in their respective zones ) has the rocket finally disappearing below the horizon 736 (well past the 535 seconds the rocket was actually under power) seconds into its flight. North Korea should have been able to receive telemetry from the rocket using land-based receivers all the way to orbital injection and even past that. (Peter has been investigating reports that a ship he believes was sent out to monitor the rocket had to return to port before the launch because of mechanical problems. It turns out, however, that such a ship was probably not needed.

A couple of thoughts immediately jump to mind:
-They most likely have not had enough time to correct what ever problem caused the April 5th failure unless it was the same type of failure that caused the August 2008 Safir second stage to fail. (I believe that the U’nha-2’s third stage is the same as the Safir’s second stage.) In that case, they could use the Iranian expertise to quickly fix the problem. But if that was the cause of the failure, why didn’t they fix it before the April launch attempt?

-The fact that they already have a rocket body ready to move to the West coast launch complex means that they are following what in the US would be called “concurrent acquisition.” In such an development path, the North would be “mass producing” U’nha-2’s even as it develops it. This has been practiced in the US, and has in fact, been used for many of the US ICBM development programs; programs which are judged so vital to the national interest that the US has been willing to run the risk of building into a number of missiles any design/production problems that might be discovered during the testing program.

-The comments above assume that simply because the rocket was visible to North Korea, the DPRK had telemetry and was able to diagnose the cause of the April 5th failure. Just because that makes a lot of sense doesn’t mean they actually had telemetry. It is conceivable that they simply have no idea why the rocket failed. (I’ve argued before that the North’s rocket scientists might not have told Kim Jung-Il that the missile didn’t make it into space. I hadn’t really thought that they might not know it failed either. However, it would not be the first time a country spent years and millions of dollars on developing missiles without any telemetry. Iraq never had telemetry on any of Al Hussein’s it fired off.)

Update: Murray Anderson pointed out a significant error I had in an earlier version of this post. Thanks! I’ve corrected that error, which I blame on my bad eyesight in confusing Isp with stage burn time in my rush to get this analysis done.

Comment [7]