Arms Control Wonk ArmsControlWonk


Last week, I was lucky to be able to join NPT PrepCom-ers in Geneva to help present the recently completed UNIDIR study ‘A New START Model for Transparency in Nuclear Disarmament,’ carried out by Pavel Podvig, Phillip Schell, and myself.

The work is accessible on the recently launched website, where you can check out the overview report, the individual country reports, some nifty maps, or download the KMZ set.

As the P5 continue to seek ways to show up on their 2010 Action Plan commitments (especially those repeatedly mentioned by name that are related to disarmament, transparency, and reporting - Action 5, 20, & 21), the aim of this project was to demonstrate the feasibility and practicality of applying New START definitions and provisions more widely to the other NPT nuclear weapon states.  Yes, New START was specifically designed for the arsenals of Russia and the United States, so why would we do such a thing?  The overview report says it best:

Even though the New START transparency and accountability provisions were developed in the context of bilateral US–Russian nuclear arms control, they could be applied to the nuclear arsenals of other nuclear-weapon states. The key advantage of New START is that it provides a legal and organizational framework for nuclear reductions that has been thoroughly tested in practice. Extending this framework to all nuclear-weapon states would be a natural and direct way of building a comprehensive system that could ensure transparency and accountability in nuclear disarmament.

Over decades of trial and error, Russia and the United States have learned many a lesson in terms of what works and what doesn’t when it comes to effective mechanisms for transparency, notifications, inspections, and general verification in arms control.  Therefore, in thinking about a transparency system for pursuing multilateral disarmament in line with the 2010 Action Plan, we think it makes good sense to build upon this experience rather than to start from scratch.

In this project, using a variety of open sources and geospatial tools, we generated New START-type MOUs for each of the P5 states, which include the relevant aggregate numbers, facilities, geographic coordinates, and weapon system technical data.

Here are the New START-type aggregate totals for each state as of 1 September 2012:

(Data for the United States and the Russian Federation come from the biannual exchange required by New START, which contained data declared current as of 1 September 2012. Data for China, France, and the United Kingdom are estimates based on open-source information.)


The numbers might look a little strange at first, but this is what you get when you apply a consistent methodology (something that the open source community tends to be rather lax about) as defined in New START.  The treaty clearly defines an ICBM, SLBM, or heavy bomber based on range, it presents good definitions for determining deployment status, and the limits of the treaty provide a useful measure for determining which forces should be considered “strategic” if only for the purposes of consistent reporting across the P5. The detailed story behind each of these numbers can be found in the country reports, but here are a few key points:

  • The only systems included in the aggregate numbers are those with ranges that meet requirements specified in New START (ICBMS, > 5500km; SLBMs, > 600km; Heavy Bombers, > 8000km or with nuclear long range (>600km) ALCMs)
  • This data reflects the status of arsenals on the date of 1 September 2012, and therefore the numbers might look quite different on another date.  For instance, part of the reason the UK’s numbers are relatively low is because two submarines (as opposed to the usual one) were considered to be in overhaul on this date.
  • The reason for the glaring zero in China’s deployed warheads category is due to the general belief that no warheads are actually mounted on China’s ballistic missiles.


The aggregate numbers in the table above do not present a complete picture of each state’s nuclear arsenal, but it’s a good starting point in terms of reporting.  More importantly, the framework and system that New START provides could of course be built upon over time to incorporate new weapon systems (i.e. Russian and U.S. non-strategic forces, French nuclear aircraft, and Chinese missiles with lower ranges) in a sustained and consistent manner.

In completing the MOUs for each state, it was also interesting how much of the necessary data was already available in the public domain.  Despite the fact that the United States removes the coordinates from its New START data releases and that Russia refuses to allow the disclosure of anything beyond its aggregate numbers, the MOUs for both of these states (including geographic coordinates for all facilities) could be rather easily assembled using open sources and Google Earth.  The same goes for the much shorter MOUs generated for France and the UK, which would each have very few facilities to report.  For China, the process required more effort, but thanks to the growing body of research and the increasing availability of commercial satellite imagery, we were able to identify many of the relevant facilities with reasonable confidence.

As an example, the map below shows nuclear forces bases and other facilities that would be included in China’s New START-type data exchange report.  Click on the image to go to the larger interactive version (tip: click on the ‘satellite’ option to be able to view the actual sites when zooming in):


With each succeeding Russian-U.S. arms control agreement, the definitions, mechanisms, and procedures have been adjusted to account for lessons learned.  New START stands as the most recent iteration of this process, and while it may not be perfect, it represents a history of cooperation and negotiated agreement between the two states with the largest nuclear arsenals.  Moreover, it contains many elements that could serve as a useful basis for elaborating a multilateral transparency and disarmament system.

As starting points go, we think that’s pretty good.

(For a more eloquent description of the project, also check out Pavel’s recent post.)

From Tomnod’s website: Examples of possible military buildings, crowds, landmarks and military vehicles in Damascus detected by Tomnod’s CrowdRank and derived from volunteer imagery insight.

* * *

DigitalGlobe announced today that it is acquiring Tomnod, a five-guy operation known for its pioneering work in the field of crowdsourced intelligence.  By combining satellite imagery and Tomnod’s own unique crowdsourcing algorithms, this group has done some remarkable work in areas as varied as hunting for the tomb of Genghis Khan, to searching for lost hikers in Peru, to tracking the conflict in Damascus.

DigitalGlobe and Tomnod have already worked together in the past, such as Tomnod contributing layers of crowd-derived insight to Digital Globe’s First Look, an online service for emergency management that provides quicker access to imagery of world disasters.  DigitalGlobe’s new acquisition of Tomnod will now result in a more sustained partnership, with Tomnod’s information being increasingly used as a data source for DigitalGlobe’s in-house analytics teams.

Here in the arms control world, the search continues (largely thanks to Rose Gottemoeller) for potential new ways to harness the information beast that is “the crowd” (e.g. tweets, photos, blog posts, media, etc.) and use it to support arm control verification efforts.  Though they haven’t engaged in this area, I think Tomnod has come up with some interesting solutions to challenges similar to those faced when considering possibilities for the role of public verification in arms control.  As some food for thought, here are a few such concerns along with Tomnod’s way of tackling similar problems:

Concern: Incentivizing people to participate in “public verification challenges” would be difficult without financial incentive (which would be too costly to sustain for a long period of time).

Tomnod’s workaround:  Make it fun. With Tomnod’s crowdsourcing endeavor to locate the lost tomb of Genghis Khan in Mongolia, they have managed to attract over 43,000 “online explorers” as volunteers which have in turn processed over 900,000 images.  How did Tomnod amass such a following of free labor?  By making the image tagging process into a learning experience that aims to feel more like a game than work.  So did they find it, you ask?  Not yet apparently, but they have reportedly made a host of other new discoveries.


Concern: Calling on the public to become whistleblowers might inadvertently put private citizens in harm’s way.

Tomnod’s workaround: Create a secure, password-protected environment where users can contribute information. Given the sensitivity of the Syrian crisis, when Tomnod worked to monitor the crisis in Damascus, it employed a “community sourcing” approach with a password-protected site sent to invite-only groups. Volunteers in the community could then sift through the images to tag things like landmarks, a crowd, or military buildings.  This reportedly resulted in a stream of intelligence tips about hotspots in Damascus, delivered in real-time to administrators and analysts via web browsers and mobile devices.


Concern: How to filter out the users who enjoy tweeting false alarms of an incoming North Korean missile attack from those who may have actually sighted something useful.

Tomnod’s workaround: A fancy algorithm.  Tomnod has developed a proprietary algorithm known as CrowdRank™ that helps sift out reliable users and reliable information from the rest of the mess.  Read more about it here.


Developing potential crowdsourcing mechanisms for arms control verification presents some unique challenges that won’t be easily resolved, but as we move forward in our thinking, it’s good to remember that crowdsourcing as a field in itself has been around for a long time and that it is well-developed in other areas and applications. I think there’s a great deal we can learn from companies like Tomnod that have taken crowdsourcing to new levels through integrating the power of the crowd with the advances of satellite imagery and remote sensing.

After all, could there really be that much difference in appearance between a camouflaged warhead storage site and Khan’s hidden grave?  The search goals may differ, but there’s much room for overlap in the mechanisms for achieving them.



Happy Friday, readers.  My name is Tamara Patton, and thanks to the ever-generous Jeffrey, I’ll be a new guest contributor here.  For those who’d like to know a little more about me and what I plan to offer, see the bottom of this post.  For the rest who just want to hear something interesting about Parchin, keep reading…

Many of you know that the alleged high explosives testing site at Parchin remains a key element of the “structured approached” that the IAEA is proposing to resolve outstanding issues on Iran’s nuclear program.  Experts in this area continue to debate whether this site is really worth the ultimatum – My colleague at SIPRI, Bob Kelley, recently gave his take in that debate, which was seconded soon after by Yousaf Butt.

As the debate continues, I thought it might be useful to take a closer look at how the many different dimensions we’ve been offered add up in the context of the site.  I came upon some interesting discrepancies.  Some of you know that I’m a fan of Google SketchUp — I gave the tool another go here.

Starting with the original drawing of the chamber (acquired by the AP and reportedly produced with information from an eyewitness), I was most struck by how off the reported dimensions seem in the context of the drawing.  Jeffrey already remarked on the oddity of the Oompa Loompa-sized door in the cylinder.  Another eyebrow-raiser is the diameter to length ratio of the cylinder (reported to be 4.6 m and 18.8 m respectively).  In fact, as far as I could tell from modeling the photo in SketchUp, the length of the cylinder in the drawing is indicative of about 6 to 7 m – nowhere near the supposed 18.8 m figure. Maybe the artist was a Picasso fan, but the disproportion in the dimensions seems a little too pronounced to have been unintentional.


If you assume a 4.6 m cylinder diameter, the cylinder length depicted in the drawing indicates 6 to 7 m, not 18.8. (Patton, SketchUp).


A related discrepancy is the placement of the collar in the drawing versus the way its dimensions are described by Danilenko (i.e. the second source of possible dimension info).  The press drawing seems to deliberately indicate that the collar is located at the rear of the cylinder.  Danilenko, however, describes that “the external part of the central section of a length of 9 m is strengthened with a reinforced concrete square section of 7.6 x 7.6 m2”.  The key word is central.  Apparently, wrapped around the center of the 19 m long cylinder would be this massive concrete collar, with the cylinder protruding on either side of it.

Like this:


SketchUp model of Parchin, illustrating how Danilenko's design would look if it were in the building. Dimension estimates: building length (30 m); building width (15 m); building height (~12m) ; cylinder total length (19m); collar length (7.6 m). (Patton, SketchUp)


Modeling these dimensions within the dimensions of the building at Parchin (heights are very estimated since these shadows are on a slope), the view above is what I got (shown with average height humans for scale).  So, is the information that came with the drawing actually at odds with Danilenko’s design?

Yes – and no.

If the length of the 19 m cylinder is partially covered in the center by the concrete collar, that leaves about 5.7 m of cylinder exposed on each side of the collar.  Interestingly, this is MUCH closer to the cylinder length that the original drawing depicts, seen here from similar viewpoints:


Comparison of AP drawing to SketchUp model depicting Danilenko’s design from a similar perspective.


So in this infamous drawing, are we maybe just seeing half the picture?  It is an interesting thought, though it seems unlikely that an “eyewitness” would have missed the other half of the 19m cylinder sticking out the back.

As far as the likeliness of Danilenko’s design being used for the accused hydrodynamic testing, there seems to be some growing agreement that the concrete collar (especially if placed around the cylinder’s center) would interfere with the devices needed to make a hydrotest test useful, such as high speed optical cameras, flash x-ray systems, and neutron detectors.  On top of this, as you can see from the SketchUp model, there’s not a whole lot of extra space in the building to support this kind of infrastructure once we use the dimensions described in the media and by Danilenko.

A final dimension of interest at the site is the heavy earth berm immediately south of the main building, which the IAEA says is indicative of the “probable use of high explosives in the chamber”.  Based on modeling off of shadows in the October 2012 Google Earth satellite image, I estimate that it is around 9-12 m tall, so very nearly the height of the building (estimated to be around 12-15 m tall).  It’s also about 6 m thick, 24 m from end to end, and located 4 to 5 m away from the main building.


The oddly shaped berm. ~9 m tall, 6 m thick, 24 m long. (Patton, SketchUp)


While this may seem like a large berm at first glance, it pales in comparison to the many other much larger berms in the greater Parchin complex.  These berms are also significantly wider, surround entire buildings, and they are sloped on both sides.  From the shadow casted by the berm in question, it is evident that it has a flat face nearest the building.  Strangely, the height to width ratio is almost exactly what is required for earth-bermed explosive storage magazines (i.e. a 3:2 slope), as illustrated in the figures in this report, though I can’t think of any sane reason for storing ammunition this close to a structure. (Ignore this! Or see the comments for some better info.)

Bob Kelley says that small, narrow berms like the one in question are more likely used to shield radiation beams like those from an industrial x-ray machine or accelerator.  He says that if the beam is collimated then you only need to shield where it is pointed.  This makes sense.  However, a remaining consideration is that unlike the closely packed buildings that the larger berms in the area are protecting, the site in question is much more isolated, and the berm is protecting the only two buildings around that would need to be protected.  I think this prevents us from completely ruling out an explosive shielding purpose for this berm.

In brief, I think that the numbers we’ve been presented in terms of dimensions for this site at Parchin bear closer scrutiny.  The two main sources of information don’t quite add up, and even if they did, the resultant design isn’t appropriate for hydrotests according to folks who know how those work.  If there are any berm experts out there, I hope you’ll give your opinion – this berm is an anomaly amongst others at Parchin, and further investigation could result in a fun new rabbit hole to dive into.

(Here’s that introduction I promised: As you might be able to tell, my favorite area of work resides at the very tiny area of overlap between the worlds of arms control and digital art.  Modeling has been very useful for wringing a third dimension of information out of satellite images and ground-view photographs, and I’m grateful to folks like Frank Pabian who have paved the way in this area .  I’m hoping to push that tool into some new directions through modeling some new sites and hardware.  I’m also excited about exploring new applications of other geospatial and data visualization tools for our field.  In short, these sorts of projects, experiments, and ideas are what I’d like to share, and I look forward to hearing what you think. -TP)