Jeffrey LewisThe Box is Back

Readers have asked me what I think of this story by Bill Broad on Syria’s decision to reconstruct a building on the foundations of the the “Box-on-the-Euphrates.”

It has a vaulted roof now, so it looks rather like a Monopoly-House-on-the-Euphrates. (MHE?)

I think Andy — aka the Nonpartisan Pundit — hit the nail on the head in the comment section when he cautioned those concluding the bulldozing of the site was suspicious about jumping to conclusions:

To be fair, let’s consider the possibility that Syria intended to rebuild this building, whatever it was. Would they not “scrape” and flatten the area to place a new foundation on?

Additionally, the two other structures in the facility, including the alleged pump-house, remain. It seems to me if Syria wanted to hide all evidence of a reactor, it would dismantle these as well. Of course that may be the intent, but something the Syrians have not yet accomplished.

Hopefully ISIS will buy future imagery of this site to monitor progress to see what happens.

There has been a certain tendency on the part of the “nuclear reactor” crowd to interpret all actions in light of that assumption (Syria must be destroying the evidence!) rather than consider alternative possibilities of the sort that Andy raised.

I still prefer the hypothesis that it is Bashar al-Assad’s naked swimming pool.

Update: ISIS does have an image.


  1. Allen Thomson

    > Hopefully ISIS will buy future imagery of this site to monitor progress to see what happens.

    ISIS does have a report on the current picture: ttp://

    A couple of points:

    – ISIS says “The height of the new building was not determined.” But both it and the previously measured (by Yale) building to the north are casting shadows in the 9 January picture. Mensuration is in order, I think.

    – The present article and the ISIS report don’t mention the 23 Nov 2007 IKONOS image that shows the new building and the new road/pipeline to the east. Somebody needs to get the hi-res version of that image (even if it means, gasp!, buying it) and see what it shows. Remember that the 24 October image showed dirt, so the new building seems to have gone up in the space of a month.

    IKONOS 23 Nov browse image:

  2. Yale Simkin (History)

    It is important to note that the Syrians did not simply “scrape and flatten” the site. They apparently shoved a major portion of the nearby hill over the area. They rushed to bury the box area, not simply level and clear. The bulldozed cover smacks of frantic concealment, not a cleaning up of a debris area.

    One possible purpose of the new structure is to conceal the lengthy and arduous task of demolishing a possible massive (and completely distinctive) foundation structures buried under the fill.

    It would be great if the IAEA could inspect the site ASAP, before it is repurposed into a LEGO warehouse.

  3. Allen Thomson

    > Mensuration is in order, I think.

    OK, I did it and, taking Yale’s value of 8 meters for the height of the “secondary structure” on the north side of the site, get 20 meters for the height of the New Big Box. I.e., not quite as high as the 24 m of BB Classic.

    Somebody should check this. Please.

  4. Allen Thomson

    > 20 meters.

    That’s at the northwest corner; the height at the ridge looks to be about 22 meters.

  5. Hairs (History)

    First of all, I’ve never written in to a website before, so please excuse me if my comments are somewhat naive.

    Regarding the height of the building: If one happens to have the time that the satellite passed over the site then given its location (and the date) some basic astronomy software will give the sun’s altitude. From the altitude (and knowing the length of the shadow) the building’s height can be calculated directly to the same accuracy as you know the length / breadth. Of course, if the ground is not level then things become more complicated – but it seems to me unlikely that a building this large (with road access) is going to be on ground anything more than a few degrees out of horizontal.

    Alternatively, if the time of the overpass is not know, one could assume that the photograph is oriented perfectly north-south and then measure the sun’s azimuth from the angle of the shadow. Knowing the azimuth, date and location one can define the time of day well enough to infer a reasonably accurate altitude.

  6. Richard F.

    Still going tall…Doing what Allen just suggested yields the same height (counting pixels) of the new building (on the lowest point of the vaulted roof) compared to the box in the august image.

    I have never seen such a tall naked swimming pool…but ice skating halls are sometimes really tall…

  7. Mark Gubrud

    What remains to be explained away, apart from the immediate bulldozing, is the fact that Syria made such muted protests to the extraordinary attack and did not invite press or the IAEA in to inspect the site even after much of the world had concluded it was actually a nuclear site of some kind.

  8. Yale Simkin (History)

    Doing a bit more shadow analysis, the MHE is a bit lower at the perimeter than the previous LEGO factory, and a tad taller along the roof peak. (I will refine the actual numbers later)

    On average, the central area which overlays the previous occupant is essentially the same height as before.

  9. Yossi, Jerusalem

    A common argument is that the Syrian Big Box must have been a nuclear facility or something equally threatening for the Israelis to risk attacking it. I think the Box may have been important to the Syrians but not necessarily an extraordinary threat to the Israelis (e.g. a workshop installing chemical warheads on improved Scud missiles). I base this opinion on a point I didn’t see mentioned.

    Syria and Iran have a mutual defense agreement, if Syria is attacked Iran is supposed to help her. Since Iran is a formidable opponent for Israel (and even the US), such an agreement gives Syria significant protection and limits the military options of Israel. If Iran will sign such agreements with other Middle Eastern countries Israel would probably feel very threatened.

    An efficient way to stop Iran taking over the Middle East by offering (or forcing) protection is to publicly show that Iran can’t deliver protection. The most direct way is for Israel to attack a country protected by Iran and get away with it.

    This seems like a good enough reason for Israel to attack Syria but what if Iran had retaliated? I guess an Iranian attack on Israel would have been a good excuse for the US to attack Iran and wipe out its nuclear facilities. Since such a counter attack may have gotten out of hand this may explain president Bush talking about being on the brink of WWIII.

    Iran and Syria choose not to retaliate openly and the head of Israeli military intelligence announced to the relevant parliamentary committee that Israel deterrence capability was rehabilitated. However, it was too soon to rejoice because for some reason Israel started expecting a surprise attack against the Dimona reactor.

    The attack on Dimona was supposed to be carried out by a commercial airliner but a massive volley of relatively accurate SS missiles (like the Syrian Scud D) saturating its defenses seems more likely. The Israelis were very worried because without this reactor their nuclear option would have deteriorated in a few years and a replacement not under IAEA supervision is probably difficult to get.

    It seems the plan to break the Iranian influence on Syria has backfired.

  10. Allen Thomson

    > One possible purpose of the new structure is to conceal the lengthy and arduous task of demolishing a possible massive (and completely distinctive) foundation structures buried under the fill.

    I agree that the purpose of the NBB, assuming that this is not all some sort of Potemkin theater, is probably to conceal further activity on the site of the OBB. Such activity being, in all probability, clean-up of the whatever-it-was that was hasitly buried after the 6 September strike.

    That being said, I remain agnostic as to whether the whatever-it-was is a reactor foundation. That seems like a reasonable guess, but we’re way out in speculation land here.

  11. Yale Simkin (History)

    “Hairs”‘s technique is a good one, but definitely a lot of effort. I hope he follows up on it. It does require knowing details of the site location and time to extreme precision. Depending on the quality of that data, the results may be either better or worse than eyeball geometry.

    The estimates used in this thread have been keyed to the estimated height of the “secondary structure”.

    It is important to note that this number was itself extrapolated from:

    1) An assumption (based on my collecting the specs on a variety of trucks used in the region) that the truck trailers on the site being about 4 meters tall.

    2) Counting pixels, which takes a bit of judging as to which are to be included

    For details see Comment 18 HERE

    Based upon the inherent softness in both the shadow measurements and the assumptions, I would say any number between 20 and 30 meters high is on the mark.

    That level of precision is sufficient to demonstrate that the building was not a low shed for storing tractors and that it is sufficient to contain a MAGNOX reactor.

  12. Mark Gubrud

    A heavy Scud attack on Dimona, heavy enough to put it out of commission for a few years, or certainly a suicide airliner (what idiot imagined that Syria or Iran would stage such a thing), would seem to also risk a catastrophic radioactivity release. But the resulting PR disaster would be the least of anyone’s worries. Obviously a major war would likely ensue.

    Assuming that The Box was actually empty or contained only uncompleted foundations for a reactor project that had been stalled or abandoned, Israel calculated that there would be no response to bombing it and little price to pay. If they thought they were risking a counterattack on Dimona and possibly “WW3”, I think they would not have bothered to bomb an empty Box.

  13. Yossi, Jerusalem

    ISIS ignores the twin site to the Big Box that is located 5 KM east. The two sites are related as can be seen by the many car tracks connecting them in Google Earth (the ground is very level and seems hard). It seems there is a permanent roadblock near the twin site (military?).

    There is a large box in the twin site that is about half the horizontal dimensions of the Old Big Box. It looks like a swimming pool but a closer look in Google Earth shows the vaulted roof. It’s interesting to note that both boxes are blue (as seen in the New York times photo but the complementary orange in the ISIS photo). A kind of international sign for a naked indoors swimming pool?

    The twin site is surrounded by grown trees and thus may be older. A reasonable guess is that the boxes are related and the increase in size hints at scaling up of the activity or the objects processed. Hey! maybe the amazing vanishing atomic reactor is actually in the twin site?

    I don’t like the idea that the NBB was built to conceal the foundations of BB. Doesn’t the IAEA have ground penetrating radar that could see the cover up?

    About the Syrians bombing a Israeli live reactor in spite of the ecological disaster. I think they didn’t really mean to bomb it, just hinting about it got the Israelis hysterical and satisfied Asad’s pride. If you allow me to be even more speculative I would say they told Israel to shutdown Dimona, thus making their threat and shifting responsibility at the same time. A wonderful conspiracy theory!

  14. Allen Thomson

    > I don’t like the idea that the NBB was built to conceal the foundations of BB. Doesn’t the IAEA have ground penetrating radar that could see the cover up?

    A box with walls and roof of sheet metal will defeat all (exterior) imaging sensors — optical, infrared, radar. You might get some indirect indications of what was going on inside, but they’d be pretty ambiguous.

    I’d guess that NBB is actually not much more than such a sheet metal shell — whether it went up in four or eleven weeks, that’s fast construction for something that size.

  15. Yale Simkin (History)

    The spa by the highway to the east appears to have no major structures taller than 10 meters or so. Both the snuffed LEGO Factory and the new Monopoly House loom way higher.

  16. Yossi, Jerusalem

    Thanks for explaining the limitations of sensor technology and measuring the spa box height.

    I understand now that NBB could be used to cover OBB remains, whatever they are. If it was a nuclear reactor with fuel rods not loaded yet, soil samples would probably show nothing also. In such case the Syrians could let IAEA inspect the place and see for themselves that Asad was right saying it was an empty warehouse.

    If it was a chemical warhead workshop maybe traces of VX/Sarin could be found in spite of all the earth moving. In such case the Syrians wouldn’t allow an inspection.

    If we further accept that the spa Little Blue Box, OBB and NBB are functionally related, maybe even scaled versions of the same facility, the nuclear theory has a problem. A half size MAGNOX reactor may not work. The chemical warhead theory fares better because LBB could accomodate a Scud in the horizontal position while OBB and NBB could take it vertically. The bigger boxes could handle better accidents like the one in A Safir, say by lowering the missile into a big concrete cylinder.

  17. Jeffrey Lewis (History)


    I would have thought that a MAGNOX reactor of any size would be taller than 30 meters.

    At least, Calder Hall and Yongbyon both are. Then again, it could be a reactor without being a MAGNOX reactor.

    Not that I think it was a reactor, just keeping the analytic possibilities open.

  18. Yale Simkin (History)


    I do not mean to imply that the existing roof height of LEGO I or MHE completes the final structure of a possible MAGNOX reactor facility.

    The boxes are appropriate for the contruction and housing of the reactor itself.

    The fuel handling hoist and related facilities would extend up further, but with a much smaller footprint (essentially an area covering the core and overlying the spent fuel and fresh-fuel handling areas) than is required for all the functions staged at the ground level.

    The Youngbyon reactor building reflects this two stage design.

    The supports for a second tier would be built after the heavy central construction is finished (they would just get in the way.)

    A second Mini-MHE would be erected atop the main structure.

    This simpler build was considered earlier by Albright and Brannan:

    The Washington Post reported on Friday, October 19th, 2007 that an official described a facility as similar in structure to a North Korean reactor. North Korean reactor construction is based on an old Russian model—in which the reactor vessel is built gradually and is not brought to the site already constructed or in large pieces, requiring a large crane to move heavy equipment inside. This North Korean/Russian approach would mean that a roof would be placed on a building earlier than in some other reactor designs, and it would hide what was inside the building earlier in the construction timeline.
    In comparing the five megawatt-electric (or 20-25 megawatt-thermal)reactor building at North Korea’s Yongbyon nuclear facility to this suspected Syrian reactor
    building, the length of the outer walls of the structures are approximately the same …
    The taller roof of North Korea’s reactor measures approximately 32 meters by 24 meters on its sides. There also appears to be a faint square on top of the Syrian building’s roof. It is unclear whether something would be built there, but its dimensions, 24 meters by 22 meters, are consistent with the subsequent construction of an upper roof.

  19. Andy (History)

    I remain neutral on the purpose of the facility given the lack of data.

    It seems to me it is a useful case study on the limitations of imagery, particularly when we only have a handful of images to work with and those we do have do not show the critical construction period.

    As a result, I think most of the theories here are plausible – even a swimming pool, though one wonders why there’s no place to get some sun!

  20. Hairs (History)

    If anyone is able to find out the time of the satellite overpass I’ll be very happy to calculate the height. It only takes a few minutes and it would be to the same accuracy as the dimensions of the sides. The alternative method (assuming north-south alignment of the photo and measuring the shadow’s angle) will have to wait until I have access to a printer, which is probably this weekend – it’s a little difficult measuring angles from a laptop screen! Either way I’ll give you my best estimate (for what it’s worth) in the next few days.

    Yossi mentions a “half size” Magnox reactor. But I think this is unlikely. A gas-cooled, graphite moderated reactor has an irreducible minimum dimension because of the very long moderation length of neutrons in graphite; it was precisely because of this very large construction size that commercial Magnox reactors were uneconomic against the more “compact” PWRs, BWRs and heavy water alternatives. Bascially, if you’re going to use natural uranium moderated with graphite (and cooled by gas) you’ll need several tens of metres of space. Of course, you could use enriched uranium (i.e. the British AGR design) but if plutonium production is your primary objective this method probably doesn’t reduce the dimensions enough to make it worthwhile. Instead I guess it would be better to use the enrichment facilities directly, and go for a uranium based weapon. A little bit of (light) water moderation, say using water as the coolant, could also help reduce the dimensions, but again you’d have to enrich. And anyway you’d basically have a Chernobyl design, with its nasty positive void coefficient at low power – leading to a tendency to go “pop” at inopportune moments.

    If the structure was / is to cover a reactor, then a small (i.e. < 50 MWth) Magnox reactor is probably as good a guess as any.

    If I may throw in a final, wild suggestion: What if the whole thing is a sham? If I were wearing the trousers in Damascus I’d be rather miffed at having had my country bombed, and I’d be inclined to rebuild the shell with a decent sized anti-aircraft capability underneath. If Israel was suspicious enough to bomb once, then they might come back for round two. At which point the propaganda value of a tracking camera showing a missile flying up into the night (even if it doesn’t hit an aircraft) is easily worth a few cheap columns and a bit of sheet roofing.

  21. Yossi, Jerusalem

    It’s a really wonderful team here!

    A question to Allen: assuming the NBB floor is a concrete cast but not too thick, is it possible to use ground penetrating radar from inside the box and see what lays beneath?

    Hairs, thanks for the info about MAGNOX reactors no down scalability. I guess this means that the Little Blue Box in the “spa” is not a MAGNOX. I understand the Norks use only MAGNOXs so if the Nork connection hypothesis is correct the LBB is either not a reactor or a great technological leap for them?

    I’m not against the reactor theory. I just feel that the big boys with the ultra high resolution photos want us to think it’s a reactor but doesn’t provide tangible proof. Why don’t they complain to the IAEA or publicly accuse Syria. Isn’t this strange?

  22. James (History)

    Hairs: Sean O’Connor’s most excellent image-analysis website focuses on locating and logging SAM sites worldwide. He’s recently taken the trouble to examine Syria’s air defense network here:

    Examining his map we find this part of Syria has no significant air defense coverage. A rather odd place to put an important strategic asset, no? But a safe place to attack a target for domestic political gain, especially if you’re willing and able to dart through Jordan and avoid the intensive SAM belts around the capital and the Golan.

  23. Allen Thomson

    > If anyone is able to find out the time of the satellite overpass I’ll be very happy to calculate the height.

    On the optimistic assumption I did it right, 08:33:30 UT (10:33:30 local) +- 15 seconds was the time. 2008-01-09T08:33:30Z for ISO 8601 fans.

  24. Allen Thomson

    > It’s a really wonderful team here!

    But of course. 🙂

    > A question to Allen: assuming the NBB floor is a concrete cast but not too thick, is it possible to use ground penetrating radar from inside the box and see what lays beneath?

    Oh yes, if you could get equipment inside the box you’d have a much better chance of figuring out what was buried even if you couldn’t dig it out (which would be my choice). Radar, acoustic, magnetic, electric resistance techniques would all be worth trying. No guarantee of success with any of them, but it would be better than being outside NBB looking at its walls and roof.

    As noted, I’d prefer shovels.

  25. Jeffrey Lewis (History)


    I know you are just trying to keep an open mind, as I think most of the readers of this blog are. I think you are doing wonderful work — much better than what passes for analysis in the chattering class here in Washington.

    I am surprised about the two stage construction method, but will look into that hypothesis.

  26. Yossi, Jerusalem

    Andy is right about the limitations of using imagery especially the limited imagery we have. Without supplementary info it’s difficult to discover what the Syrians did in the Old Big Box. In such a case it may be helpful to solve an easier related problem: what the Israelis thought about OBB purpose.

    Well, ACW agent in Jerusalem found a newspaper article written by Ron Ben-Yeshai, the legendary military reporter of the largest newspaper in Israel. Ben-Yeshai is well known for fearlessly reporting from Arab countries and other hostile places.

    The 13.1.2008 article looks like a speech by a military photo analyst that has great doubts about Syrian honesty and integrity. In fact he accuses them of trying to cheat a future IAEA inspection. I’ll try to summarize his claims:

    * there are canals, probably new, from NBB to the river
    * OBB was covered with dirt and rocks that were leveled
    * NBB is the same size as OBB
    * OBB didn’t have a roof so one could see it had several floors.
    NBB have a vaulted roof so you can’t look inside
    * photos made about a week ago (ISIS?) show an orange colored roof.
    photos taken a few days later show the roof painted in deep blue
    * the orange is anti-rust base paint. the deep blue makes NBB look
    non-military at night
    * OBB had a small concrete wing attached to it, NBB doesn’t
    * it took more than half a year to build OBB but it wasn’t
    finished (it didn’t have a roof yet). NBB was built,
    including a roof, in 2 and half months
    * there is no sign of concrete casting in the site so NBB
    is not made of concrete like OBB and is a quick steel job
    * NBB is a closed empty metal frame used to hide evidence
    removal activities going on underneath: soil purification
    and removing OBB remains. the Syrians may be digging
    radioactive materials or fine machinery connected with
    fissile materials (centrifuges?)
    * OBB was made of reinforced concrete and had several floors.
    NBB is just a high shade

    It seems the mystery analyst is very concerned that IAEA inspectors wouldn’t be cheated by the double crossing Syrians. An interesting policy change after refusing to cooperate with them and supply info.

    I think this analyst doesn’t have good photos or doesn’t know how to interpret them. I wonder if the air strike was based on lousy intelligence. Please correct me if I’m wrong.

  27. Yossi, Jerusalem

    I beg the anonymous analyst pardon, his claims (except OBB/NBB size equality) may be correct assuming the Syrians removed OBB roof about half a year before the air strike and didn’t put it back. In such case Israel and the US could peek inside while construction was taking place. Maybe they saw the concrete cylinder reported by the Mossad spy?

  28. Hairs (History)

    Thank you very much to Allen Thomson for posting the overflight time – I believe it is accurate, as you’ll see in a bit.

    Given a location of 35 deg 42’ 28 sec N, 39 deg 49’ 59” E (from Google earth), and knowing the overflight was at 08:33:30 UTC, I get the sun’s altitude as 26.850 deg and the azimuth as minus 14.022 deg. If you want to check this you can use the same site as I do: Alternatively any other reputable site, software or almanac should give the sun’s declension and hour angle / right ascension, and then you just plug in your latitude and longitude to get altitude and azimuth. Note: unlike a bearing, azimuth is measured from the south with positive values towards the west. Thus the sun should have been about 14 degrees east of south, which would cast a shadow 14 degrees west of north. As a check, I’ve managed to get a print of the ISIS pdf file: the pixels are a little rough, but by protractor I measured a shadow angle of 14.5 degrees, and with a ruler (and a little basic trig.) I got 14.15 degrees. So all in all the assumption that the photograph is aligned north-south seems justified (useful for any future analyses perhaps) and I’m confident that the overflight time, and the sun’s altitude derived from it, are valid.

    As to what all this tells us: if the sun’s altitude was 26.850 degrees then the height of the north west corner of the roof is 0.235 x the width (i.e. the east-west dimension) of the building. Given the ISIS estimate that the building is 60 metres by 60 metres, this puts the edge of the roof at 14.1 metres high. By a similar method, the raised centre of the roof is then 15.7 metres high.

    I know this is lower than some other estimates, but there’s certainly no doubt about the site’s location, and the overflight time seems reliable enough. It is possible that the shadows are not cast on level ground, but the general straightness of all the shadows, plus the fact that Google earth doesn’t show any significant changes in elevation where the shadows fall, suggest that there’s nothing that could cause an error of, say, 10 metres in the calculation. This leaves only the length / width of the building itself: I’m taking on trust the ISIS estimate of 60 m x 60 m, but presumably these people are no fools and thus have a reliable scale indication for their photographs.

    I’m reluctant to say that ca. 15 metres high is the “right” answer – I’m way out of my field here, and could be making all sorts of errors and incorrect assumptions, but I do think 15 m is a lot more plausible than, say, 30 m. This building is BIG by any standards: even if the roof is just a bit of cheap 2 mm corrugated sheet it’s still 7 cubic metres of metal, which is something just over 20 tonnes if it’s made of aluminium! (Actually it’s more because of the corrugations). Then there is wind loading, and seismic loading (our revered leader would be irked no end if a minor tremor brought the roof down on his plans for world domination). Oh, and we mustn’t forget the clearspans. If you’re going to build a reactor in the middle then you can’t very well have a stray steel column passing through its centre. All of which means the roofing spans are going to have to be strong enough to stretch a good 20 metres or so without support, and still have enough strength to carry their own deadweight, plus that of the roof, plus the dynamic loads. Those are big lumps of metal. And then don’t neglect the lifting provision: if you have a working reactor you won’t exactly have room to drive a mobile crane inside when you want to lift a pump, or pull a fuel bundle; so you’ll probably want to either hang a traveling crane from one of the spans (more dynamic load) or else run a gantry crane atop your columns, which, you’ll be relieved to hear, finally brings me to my point: Columns!

    The kind of columns that you need to hold up a 60 × 60 × 30 metre building, and then hang some lifting provision off of, are going to be long and heavy. I’m not a structural engineer, but (in seismic zones) turbine halls with gantry cranes considerably smaller than what we are looking at here are often built with W36 × 300 columns. That means a column 36 inches wide and 300 lbs per foot of length! So a 30 metre high building is going to take columns 100 feet long and 30,000 lbs in weight. Even if I’m generous (miserly?) and build my MHE with W18 × 130 columns, I’ve still got to handle something 100 feet long and 13,000 lbs in weight, and do it many times over in a relatively restricted site. If you look at the roads in the photo I just can’t see a turning circle for a 30 metre pre-fabbed column. And even if they fabricate on site, it’s not your average crane that’s going to handle 15 tons, 100+ feet in the air and boomed out tens of feet away. I’m not saying all this can’t be done – a building is plainly there, and even 30 metre high giant buildings are constructed around the world; it’s just not done easily, cheaply or without good reason.

    So we’re back to two choices:
    1) It’s the mother of all buildings, constructed surprisingly quickly at a remote(ish) site in a country that doesn’t rate high on my list for engineering infrastructure and support services (read: big, mobile cranes). Thus by its very size and speed of construction it MUST be for something important.
    2) It’s a largish building – there’s no denying the area of roof – (elder sister of all building, perhaps?), that could have been built comparatively easily by bog standard mobile cranes (a 15 metre long W18 × 130 is handled easily by a normal truck and mobile crane) and which may, or may not, be important.

    Personally I go for option 2, sub-section “not important”. The more I think on the idea of it being a cheap lure to invite another air attack the less crazy it seems! Thank you here to James for the link to Sean O’Connor’s amazing website about SAM systems – I thought I was the only one with unusual interests!

    Sorry to all if I’ve needlessly consumed your bandwidth; I’ve been carried away in the novelty of seeing something I’ve written appear in “public”. I promise to be more restrained in future.

  29. Jeffrey Lewis (History)


    Don’t be more restrained.

    The whole point of having comments on the blog is to provide a forum for interested folks with real skills (i.e. not looking pretty and talking nice) to discuss security issues with an open mind and a ready calculator.

    This is great.

    One question, are you measuring from the corners or the apex of the vaulted roof?

  30. Hairs (History)

    Thank you for the kind encouragement – but I really have no qualifications or experience in this type of business, and I didn’t want an excess of exuberance to offend anyone.

    Regarding your question: if it’s about the heights, I make it 14.1 m from the ground to the north west corner of the building, and 15.7 m from the ground to the apex. If your question was actually about the azimuth angles (i.e. my check of the overpass time), the truth is I measured all three angles: NW corner to NW shadow, apex to apex shadow, and NE corner to NE shadow, and took an average.

    One final point, I should perhaps correct my “not important” statement: What I meant was that I don’t think the building is important in and of itself. It surely can’t be to fulfill the same role as the original structure because in that case the original was 13m too small on all sides (which would be a distinct embarrassment if our beloved leader’s doomsday machine was supposed to fit inside!). Therefore it seems to me that the new building must be a “quick and dirty” structure, slapped up in just 10 weeks, and specifically sized to be bigger than the original; and in my mind the only serious reason for doing that would be if the new structure is there simply to hide the rubble / deconstruction of the old structure. As to the height, 15m is probably necessary if you want to get some decent sized plant inside (e.g. coring machines, cranes, etc) to remove any deeply set rubble. As I say, I don’t believe the new one is important, except inasmuch as it is hiding the old one.

    The one new piece of knowledge that does seem to be indicated by the new building, is that the old structure WAS important! Although what that was, or was intended to be, I really have no idea.

  31. Mark Gubrud

    The new box is not the old box. It was constructed too quickly to be anything more than a roof over the area. Those of us who think this was probably an uncompleted reactor project are guessing the purpose of the new box is to prevent observation as the site is excavated and foundations are dismantled. Another possibility is that if something chemically toxic or radioactive was present at the site and has been buried, the new box may also serve to contain it as the site is excavated. In any case, the new box is apparently high enough for excavation work to proceed underneath and its height does not tell us anything about the height of whatever was or was not yet inside the old box.

  32. Yale Simkin (History)

    If “Hairs” analysis is correct, then it is much less likely that a MAGNOX-type reactor was being built on the site.

    Comparing shadow lengths between the “secondary structure” and the Big Boxes in both the old and the new images, shows that the current building is essentially the same height as the destroyed earlier building (altho I need to check some other images).

    If the height is scaled down from my estimated 24 meters to Hairs’s 15 meters, then other items at the site must also be downscaled.

    Of most interest is the downscaling of the trucks on site. The cabs and trailers would drop to only about 2.5 meters ground to roof. However, in the images, measuring the shadows of those trucks is rather tough due to pixel size.

    I will replicate Hairs’s method and report back.

    Again, assuming his derived numbers are good, then a MAGNOX possibily is severely limited.

  33. Rwendland (History)

    Early commercial Magnox reactors had concrete biological shields about 3.2m thick . In a military environment maybe a 2m thick shield would do. Is it realistic that a few Israeli planes and whatever equipment the Syrians got there could demolish something so substantial in a few weeks?

    The core would probably need about 200 tons of graphite (50MWe Magnox used 1140t ). Because of access issues, I’m pretty sure the graphite core, and quite a lot of other core equipment, would have to be assembled before the pressure vessel and concrete biological shield could be finished. So if we were seeing the finished concrete shield of a reactor, I suspect a lot of the reactor would have had to have been completed. Though no sign of a spent fuel pond that would be needed. And there is probably a lot of graphite to be found in the rubble.

    Have to say, the reactor theory doesn’t strike me as all that plausible.

  34. Yossi, Jerusalem

    Yale, I suggest a method that seems safe to calculate a box height from its shadows using a pre-calculated sun altitude. I’ll present it in stages:

    Let’s calculate first the height of a vertical pole. Assuming the sun altitude, let’s call it theta, is measured from the horizon and up to the zenith we have:

    tangent(tetha) = pole_height / shadow_size

    pole_height = tangent(tetha) * shadow_size

    Now let’s calculate the height of a flat roofed box:

    * choose, if possible, a vertical edge on the shadowed side. You should be able to locate its two vertices, i.e. the roof vertex and the base vertex. If you don’t have such an edge then there is no deal. It should be possible in half the cases, on average.

    * find the shadow of the roof vertex on the ground

    * measure the distance from the roof vertex shadow to the base vertex

    * substitute the measured distance in the formula above in place of shadow_size

    * compute the pole_height, that’s the answer

    A bi-inclined roof is a more complicated case. If roof inclination is small you will be able usually to calculate the height of a low side. To compute an “apex” you will have to locate its corresponding “base vertex”. If the two sides of the roof are equal the “base vertex” will be a midpoint of the horizontal edge below the apex.

    Too complicated?

  35. Allen Thomson

    Here are my estimated times for other Digital Globe imagery that shows shadows cast by OBB pre-strike and other things in the 24 October post-strike image. I encourage the umbrologists among us to see if anything interesting can be learned from them.

    DG 5 August 2007 , Appendix B
    Estimated time of image: 2007-08-05T08:36:20Z

    DG 10 August 2007
    Estimated time of image: 2007-08-10T08:41:30Z

    DG 24 October 2007
    Estimated time of image: 2007-10-24T08:27:20Z

  36. Yale Simkin (History)


    Could you recheck the times you have associated with the images? The solar angles don’t seem quite right on the 2007-8-10 image.

    Where did you obtain the metadata?

    I want to make sure the numbers are right for the old LEGO Factory.

    I did some “umbrology” on the new MHE and they agree with Hairs.

    Here are my tentative results:

    For the MHE image of 1/9/08 and using the 62 meter roof edge size, I got the same result as Hairs, 15 meter height for the NW corner. (I used 30.726 deg for the Sun altitude – Hairs, did you use Mercury’s altitude instead of the Sun’s?)

  37. Davey (History)

    This is a facinating peek! I’d appreciate it if you would clarify something for a non-expert. Is there consensus that there is quite a bit of fill and construction debris covering the construction site? (Yale’s first post) If so, there is a fair chance that quite a bit of construction is below grade. They’d have to remove all of the uncompacted fill before construction, so why not excavate a little deeper? If the building is just a free-standing shell, it can be built at ground level on concrete caissons. This means that the building can be lower if there is a significant ‘basement’.

    The sheer amount of earth moved already makes it difficult to detect any new fill on the site attributable to excavation within the building. Anybody remember watching the movie “The Great Escape” – how American POWs disposed of excavation material from their tunnels?

    Just a couple idle thoughts from an engineer on lunch break. Thanks!

  38. Andy (History)

    Davey brings up a good point which would tell us a lot about the purpose of this new building. If the foundation for the NBB is essentially perimeter piers with a clear span interior then that would provide the possibility of covert excavation of the OBB, particularly since this new building is wider and shorter.

    On the other hand, if there’s a poured concrete floor with a lot of support columns on the interior, that would indicate another purpose since interior support columns make excavation tricky at best. The later type of structure would be easier to build (think of a super Wal-mart for example) than the former since minimizing interior columns is tougher.

    The presence of a peaked roof instead of a sloped or flat roof indicates the roof structure may be trusses which is what one would want for minimizing or eliminating interior columns.

    It’s really too bad we don’t have an image showing this new building mid-construction – that would tell us an awful lot.

  39. Yale Simkin (History)

    The material that appears to have been bulldozed over the site is seen HERE

    The new box appears to be on a raised platform with a berm or retaining wall to the south – HERE

  40. Allen Thomson

    > It’s really too bad we don’t have an image showing this new building mid-construction – that would tell us an awful lot.

    Which is why it might be useful to get the 23 Nov 2007 IKONOS high-res shot. To my eye, the browse image indicates the roof was already up (there seems to be a shadow on the north side), but you go with the data you have.

    Yale: I’ll recheck the pass times for the imagery tomorrow; life intervened today. The metadata are orbital elements of the satellite for the days in question run through a vanilla prediction progam, Traksat 4.09 in this case.

  41. Hairs (History)

    Oops! Yale is quite right – I must have accidentally used Mercury’s altitude instead of the Sun’s. Sorry for that, and thank you very much for checking!

    If I correct my original estimates, I end up at 16.5 m for the corner and 18.4 m for the roof apex, which in the grand scheme of things is the same as Yale’s values. After all, the pixels are a bit blurry, and at the angles we’re talking about a change of (an assumed) building width from just 60 m to 62 m will change the height calculation by 0.5 m.

    I haven’t had an opportunity yet to look at Allen’s links for the old building, but I’ll post any results I get as soon as I can.

    Allen – I love the “umbrology”. May I suggest that you even re-name the whole field of arms control wonkery as “penumbrology”, since it’s permanently located between the shadow of secrecy and the illumination of open, public knowledge! :-))

  42. Yossi

    There is another absolute method to measure heights, using relief displacement. In some BB photos the displacement is quite large, more than the shadow and so can provide more accurate results. The formula is simple:

    h = (H*d)/r


    h – object height

    H – height of camera

    d – the relief displacement

    r – distance of object from nadir (if photo is vertical, nadir = photo center)

    If you have complete photos and can assume they are vertical you only need the satellite height above ground, everything else can be measured directly.

    Can you please explain why in Hairs’ method box height equals a constant times box WIDTH? How do you get this constant?

  43. Allen Thomson

    > Could you recheck the times you have associated with the images? The solar angles don’t seem quite right on the 2007-8-10 image.

    OK, I checked them and get the same numbers. It’s possible that there might be one minute error, maybe a little more (though I don’t think so) if the picture were taken away from the point of QuickBird 2’s closest approach to OBB.

  44. Hairs (History)

    First of all my apologies for posting this very late – I haven’t had any access for the last week so I’m afraid I’m a bit behind.

    I have had a look at the Digital Globe image at , Appendix B (thank you to Allen Thomson for the reference) and concentrated on the rather fetching oblique view of the “Old Brown Box” on page 313 of the pdf file.

    According to Allen, the image was made at 08:36:20 on 05 August 2007, which means the sun was at an azimuth of minus 34.195 deg and altitude of 68.235 deg. Now, in the image the shadow at the north west corner of the building appears to lie almost parallel with the northern roof edge, thus if you print the image and measure the angle from the north west corner of the roof to the north west corner of the shadow it will appear that the sun’s azimuth is something around minus 90 degrees. However, this is just an artefact of the image caused by its being taken at an oblique angle. (Yale – could this be the source of the mismatch between your angles and Allen’s overpass time?). Instead, in order to estimate the true azimuth in the image, draw a short line “down” the corner where the eastern and southern walls meet i.e. imagine you are standing on the roof at the south eastern corner and you drop a thick plumb line (thick enough to be visible in the photo) down to the foot of the corner. Now parallel translate this line to the north western corner of the building. The “foot” of your line is, of course, hidden in the shadow, but it represents ground level at the junction of the northern and western walls. NOW draw a line from the foot of the north western corner to the north western corner of the shadow, and measure the azimuth angle. This is the true azimuth of the sun in the image. With a little trig I get minus 33.2 degrees, and with a protractor I estimate minus 32.9 degrees. Consequently I think Allen’s estimate of the satellite’s overpass time is correct (after all, orbital calculations tend to be a bit more accurate than me gazing bleary-eyed at a protractor over a grubby black and white print-out!).

    So, if we assume that the image WAS taken at 08:36:20 that means we can go with an altitude of 68.235 deg. Since my ruler gives me: 15mm from the “foot” of the north western corner to the corner of the shadow, and 92mm for the building’s east-west dimension (I prefer east-west because it reduces any error due to foreshortening when measuring north-south), I can then use the published assumption of a 47m building width to end up with a shadow “length” of 7.66m. Finally, with a length of 7.66m and the tangent of 68.235 degrees I end up with a building height of 19.2 metres. Yossi – I hope my long-winded explanation addresses your query on how I leap (to the conclusion of) tall buildings from the single bound of the building width. 🙂

    As a plausibility check on all my shenanigans I thought I’d try to bracket the possible heights that could be obtained from the photo. Knowing that the image came from the Quickbird satellite, and with a little bit of help from my friend the internet, I find that Quickbird’s swath width at nadir is 16.5km. Furthermore, its manufacturers say that its field of view is 2.12 degrees, and perigee and apogee are almost identical at 445km and 447km respectively – which very helpfully works out at a swath width of 16.504 km. So, a swath width at nadir of 16.5 km it is then! Additionally, I also find a reference to another swath width of 544 km at ca. 30 degrees off nadir.

    Next, a quick review of “Orbital Mechanics for Dummies” tells us that an orbit of almost zero eccentricity (thank you Ball Aerospace!) at 446 km altitude has an orbital speed of about 7.7 km/s. Thus Quickbird passes a 16.5 km swath at nadir in about 2 seconds, and even a 544 km full swath off-nadir is passed in about 70 seconds.

    Well, from the image the photo was clearly taken well the building was well off nadir (otherwise we wouldn’t see the sides so well) so, as an easy approximation, let’s just say that the photo could have been taken anywhen within a two minute window (one minute either side of Allen’s overpass time). Thus the earliest the photo was taken was around 08:35:20 and the latest was around 08:37:20.

    I won’t bore you with the maths again (though please feel free to check – after all, I got it wrong earlier) but changing the photograph’s time by +/- one minute changes the height by about +/- 0.1 metres.

    In short, my calculations suggest the OBB’s height was 19.1 – 19.3 metres. Of course, given my dodgy methods for the length of the building’s shadow, I’d freely give +/- 5 metres error on the final answer, but overall I reckon the OBB was about 20 metres tall, and certainly not much more than 25 metres.

    I’ll have a go at other photos when time allows, but in future I’ll just post my results unless anyone would specifically like to see the method / workings.

  45. Hairs (History)

    Regarding Yossi’s comments on relief displacement, I agree that in principle it would be a more accurate method of analysis. However, because we are dealing with satellite imagery and its associated “high” camera heights (rather than the more modest camera heights of aerial photography) this method becomes very sensitive to the distance between the object and the nadir.

    For example, if we go back to the oblique photo mentioned in my previous post, the camera height is known to be 446 km. The displacement (i.e. the distance in the image between the foot of the southern wall and the roof) is about 9.2 metres. And the distance from the nadir is unknown.

    Clearly the image cannot be at the satellite’s nadir – otherwise we wouldn’t see the southern wall – so I’ll start by assuming that the building was right at the edge of the “nadir image” swath i.e. 8.25 km from nadir. This gives a building height of 497 metres, which is probably a little flattering to Syrian construction capabilities!

    At the other end of the scale, if Quickbird 2’s maximum image swath is 544 km wide, that means the building could have been 272 km from nadir when the image was taken. (Personally I think this distance is unlikely because the resolution is still pretty good compared to known “overhead” photos). Anyway, at 272 km off nadir the building’s height becomes abut 15 metres. Similarly at 205 km off nadir its height becomes 20 metres, and at 164 km off nadir its height becomes 25 metres.

    In other words, until or unless we know the exact nadir of the satellite at the time the photo was taken (e.g. from a time stamp on the image itself) any plausible building height – as well as quite a few implausible ones! – is possible.

  46. Hairs (History)

    After mulling things over, I’m pretty sure I should not be ignoring the 05 Aug 2007 photo’s foreshortening in the east-west direction. Therefore I’ve had another bash at both methods (sun’s altitude and relief displacement), this time using the dimension across the diagonal (which should be 66.47 metres for a 47 m x 47 m square).

    Cutting to the chase, my revised estimate (guess??) is an OBB height of 26.1 metres, with an error of at least +/- 5 metres.

    Sorry this doesn’t advance us very much – I mean, we already knew it was going to be between 0 and 50 metres high! – but it suggests that the OBB probably WAS tall enough to house a small gas-graphite reactor (if that is indeed what it was for). Even if it wasn’t for a reactor, I’m confident that the building was significantly taller than 15 – 20 metres, which unequivocally means that its height was an important feature; so a tractor shed it is certainly not, although indoor swimming pool is still an outside chance if our revered Father of the Nation is secretly practising for cliff diving at Acapulco…

    A best-guess height of around 26 metres also tends to confirm the impression that the “new box” is substantially lower than the “old box” – implying that the new box is NOT a replacement for the old one, but instead must have some other purpose.

    Using relief displacement, a true height of 26 m would mean that the OBB was 205 km from the satellite’s nadir when the photo was taken. But beyond that I can’t really refine things much.

    If anyone can find nice oblique satellite photo WITH A TIME STAMP I’d be very much obliged! :-))

  47. Yossi, Jerusalem

    Hairs, you did a wonderful work! Thanks for the excellent explanation!

    * We poor photo gazers have difficulty with parallel transporting of box edges without some equipment. I wonder if it would help to have software that you feed (using a mouse) with those box/shadow corners you can see and it gives you those you can’t?

    I think that such software wouldn’t be too difficult to write. In a photo of a rectangular box there are three rectangles: the roof, base and shadow. In general only the roof will be seen completely, the base and shadow will be partially obscured. Since satellite height is so much larger than most ground details the rectangles will be translated relative to each other but be practically equal.

    * In the relief displacement formula above “r” should be the distance of object TOP from nadir, not the base line distance.

    * If OBB height was 26m it could accommodate a MAGNOX reactor, especially if they digged a bit inside. It could also be useful for erecting a missile that is longer than a Scud/Shihab inside a cylindrical shield and work on its chemical warhead.

  48. Yossi, Jerusalem

    I seem to be endlessly correcting my own posts…

    * Fitting a chemical warhead to the Syrian missiles proved problematic and probably made A Safir, their main base, unusable for a long time. The simplest solution is to lower the missile vertically into a hollow cylindrical shield that is higher than a missile length and work on it inside. In case of accident you have to quickly flood the cylinder with water that will combine with a few barrels of NaOH. Scuds are about 11m long so you need at least a 22m roof. A safety margin and the distance between the crane system and the roof can well add the total height to 26m. OBB seems the perfect site for such work because it’s hidden in a creek, near a river and railroad and has a sister site suitable for the staff.

    * The Israeli prime minister gave a few days ago a hint about the mystery air strike. He said that even if you are not a party to the secret you can see that our enemies in the North are not eager to attack us. Remember that the head of military intel said after the strike that Israel’s deterrence capability was restored. Now note that Syria has a mutual defense agreement with Iran so any attack on Syria is a test of this agreement. Attacking Syria with no Iranian military response shows the agreement is useless and Iran is a paper tiger.

    I think OBB was not a significant threat to Israel, it was attacked as a test of the Syria/Iran agreement. It was a political gamble that could lead either to war or a demonstration that Israel is still the top dog (when backed by the US).