Jeffrey LewisIranian Nuclear Test Shaft

Earlier this week, Dafna Linzer had a long, detailed story (and online chat) on intelligence found by German intelligence on a laptop computer stolen by an Iranian citizen in 2004. I am going to write a four-part series looking at the key pieces of information: a schematic of a shaft that might be for a nuclear test, plans for an underground facility to produce uranium tetrafluoride, modifications proposed for the Shahab 3 ballistic missile and the acqisition of relatively advanced P2 centrifuges from Pakistan.

Dafna Linzer’s article on Iranian nuclear activities opened with the revelation that Iran might have a diagram of a “400-meter tunnel” that inclided that a senior US source described as “consistent with a nuclear test-site schematic.”

The details—the depth of the shaft, the existence of diagnostic equipment, and the existence of a control facility a few kilometers away—are consistent with a nuclear test site. Still, I am not sure why a diagram of a nuclear test site wouldn’t be immediately obvious.

[Update, I’ve posted a link to a DTRA brochure explaining how to distinguish nuclear explosive testing from legitimate mining and excavation activities.]

A new nuclear state might attempt an underground nuclear test—which is harder than just blowing the bomb up above ground—to contain radioactive fallout that might threaten neighboring populations and, perhaps more importantly, provide intelligence information to other countries that collect the debris.

The depth of the shaft, 400 meters, is basically the correct depth to contain an underground nuclear test of a simple fission-type nuclear device, using the US formula for such things:

Depth, m = 122 (yield, kt)1/3

I should add that Linzer doesn’t make clear whether the schematic depicts a vertical shaft (right) drilled into the earth or a horizontal tunnel drilled into a mountain (at the beginning of this post). Linzer uses both “tunnel” and “shaft” to describe the diagram. The difference is not crucial—although one might imagine other uses for a tunnel dug into a mountain.

Linzer describes “remote-controlled sensors to measure pressure and heat.” These would, I assume, refer to diagnostic canisters (pictures) that contain sensors to measure the explosion.

It isn’t clear to me why the schematic or diagram would not be obviously associated with a nuclear test site—I wonder about details that might suggest another purpose for the shaft. Linzer notes that the US has “no evidence of an associated program to acquire, assemble and construct the components of such a site,” suggesting that such a schematic might be a preliminary investigation.

An Iranian asked to look into how Iran might construct a test shaft could draw on a wealth of information including US publications like: Office of Technology Assessment, The Containment of Underground Nuclear Explosions, October 1989 and Vitaly V. Adushkin and William Leith, The Containment of Soviet Underground Nuclear Explosions, September 2001.

Iran might also have access to information from Pakistan—although Khan was not in charge of the Pakistani nuclear test and his deals with Iran predate Pakistan’s nuclear test in 1998.

That’s nuclear entrepeneur AQ Khan, at right, in the Pakistani test tunnel.

Pakistan used a horizontal tunnel for its 28 May 1998 nuclear test(s) and a vertical shaft for the 30 May 1998 nuclear test(s).

A Pakistani journalist reported that Pakistan’s May 28 event was conducted in horizontal tunnel with a 700 m overburden and an observation post about 10 km away—essentially a similar set-up to the schematic that Linzer describes in the Post.

Altogether, the description of a test shaft seems plausible given the details. Still, I have this nagging doubt about why the diagram’s purpose isn’t a “slam dunk,” if you’ll pardon the expression.

That slight sense of doubt is exacerbated by the decision to leak this information. If this really was a diagram for an Iranian nuclear test site, then the US IC would use information in it to look for signs of an Iranian nuclear test down the line (see, for example, Vipin Gupta and Frank Pabian, Investigating the Allegations of Indian Nuclear Test Preparations in the Rajasthan Desert, July 1996). Presumably, the Iranians would alter or camouflage any test signatures revealed by the diagram—increasing the chance that they might be able to evade US intelligence satellites as India did in 1998

Of course, the diagram might be a lot less detailed than that, which leads me back to my nagging doubt.


  1. Cheryl Rofer (History)

    Thanks for going into the wonkish details, Jeffrey. As I said when the Linzer article came out, it’s unsatisfactory in that it pretends to detail, but when you try to do this kind of analysis, the detail really isn’t there, down to whether we’re talking about a shaft or tunnel (adit is the correct mining term).

    If this is indeed real data (and, like you, I question that), I would guess, because mountains are mentioned, it’s a tunnel.

    As to reasons for an underground test, I’ll add another: ambiguity. That ambiguity can be useful in case the test doesn’t go well, as is thought to be the case for the Pakistani tests. It can also be useful in allowing the pretense (by Iran or by others during negotiations) that we don’t know that it was a nuclear test.

  2. Pavel Podvig (History)

    I would say it doesn’t make a lot of sense to start digging a tunnel when you don’t have material for a bomb yet. It would be a rare example of advanced planning.

  3. Jeffrey Lewis (History)

    Again, no digging. Just a schematic that, as far as I can tell, is not linked to a specific place.

  4. Arrigo (History)

    I was leafing through my beloved copy of Dahlman/Isrealson “Monitoring Underground Nuclear Explosions” and couldn’t find any specific reason for looking into horizontal tunneling as opposed to vertical tunneling except for geology and/or experimental requirements (like Teller’s laser stuff, I imagine).

    Why would Iran go for a horizontal tunnel instead of a vertical bore given all the oil industry experience they must have in going down vertically?

  5. Jeffrey Lewis (History)

    Horizontal tests are easier to instrument. Here is how the Office of Technology Assessment (1989) described the difference (page 18):

    Presently, an average of more than 12 tests per year are conducted at the Nevada Test Site. Each test is either at the bottom of a vertical drill hole or at the end of a horizontal tunnel. The vertical drill hole tests are the most common (representing over 90% of all tests conducted) and occur either on Yucca Flat or, if they are large-yield tests, on Pahute Mesa. Most vertical drill hole tests are for the purpose of developing new weapon systems. Horizontal tunnel tests are more costly and time-consuming. They only occur once or twice a year and are located in tunnels mined in the Rainier and Aqueduct Mesas. Tunnel tests are generally for evaluating the effects (radiation, ground shock, etc.) of various weapons on military hardware and systems.

  6. Arrigo (History)

    True, they are easier to instrument but my point was that as a first test is it also not much harder to put together a safe horizontal test than a nice deep vertical bore?