Rooftop Impact

Ghouta Chemical Massacre:
Firing Directions 
Part 3: Ain Tarma Rooftop Impact
August 11, 2017
incomplete
edits Aug. 12

adds 3/28/21

This is one of three Zamalka-Ain Tarma impact sites inspected by the UN-OPCW investigators on August 28. It was investigated at the same time as impact site #4 (see part 1), apparently a few minutes earlier (all between 2 and 3 pm local time). As they saw two spots previously in Moadamiya (August 26, listed as sites 1 and 2), this roof area must be impact site #3. (and #5 is an unplaced home, numbered impact 7 in my roster, which they visited on Aug. 29)

In the views at right (from this video), we're inside a room at roof-level, looking west or northwest across this and other rooftops. The rocket apparently blew through this west wall, and smashed through the reinforced concrete floor just inside, on roughly the camera's line of sight. It's also, apparently, sort of pushed out a mostly intact south wall panel, which seems strange (and so we're looking at an angle through two breached walls). To me, this scenario makes sense, although I'm no expert, and questions can be raised. We'll come back to these impact details.

Analysis of the actions in this video seems worthwhile for an expert - it seems the investigators warn the rebels off the roof, based on dangerous reading they're getting of ... something, here about a week after any surface or floating sarin should have been evaporated away. Is that perhaps suspicious, suggesting it was planted more like the day before? He seems to show them his reader to emphasize the point, and the less-Islamist looking "FSA" guy makes a show of covering has face off and on, and trying to shoo the cameraman away. The more bearded one seem completely unworried, just watching the investigators collect their samples.

Another video shows the OPCW crew in the room below examining the twisted remains of the rocket next to a coffee table. That's less clear and harder to correlate, so it won't be analyzed here. The state of the rocket may be of some interest, but I've nothing to say on it right now.

However, briefly, here's a new tidbit: researcher Chris Kabusk tweets that a photographer "who went there" later (for the photo in 2016?) and that that he's in contact with " also found the rocket sitting just below the rooftop, still." He provided this comparison of 2013 video stills and a Reuters-stamped photograph I've never seen. It shows a red rocket ID#, only the second one I know of from this incident: #153. (

8/13 note: one other #165 has emerged, but location unclear)

But back to video analysis and geolocation, and then angle of fire.

Geolocation
The site was initially hard to place, with little visual information to go on. Researcher Chris Kabusk had in 2014 decided on a spot that now seems incorrect. Coordinates: 33º31'17.5” N, 36º21'8.2” E (Mapped with 3D analysis here, along with a nice 3-D model of the impact site to help us visualize). It seems possible he took a shortcut here and took one of HRW's identified spots (on the map below, the one closest to "garage door"), and found a roof very nearby with a little structure in the middle he took to be this one. That's reasonable enough as a guess, but for one thing, I noticed right off (when I caught it recently) that the surroundings never did match, even the immediate rooftop shapes and details.

The UN's report (PDF via publicintelligence.net or UN PDF ) specified the site on a roof of a 5-story building was visited by one team (samples taken from 2:14-2:40 pm on the 28th) at the same time another investigated impact site #4 in a "nearby open field," taking samples from 2:34 to 2:51 pm. (see page 25-26)  When Kabusk returned to this in recent weeks, he found a 2017 photo of the same site, said to be in Ain Tarma  (where the field impact was), not in Zamalka. (this is included in the graphic below).

We all looked for a match, searching for similar roof shapes exceptionally close together, and locating a nearby minaret and a distant mountain, to find a basic line-of-sight. But in the end it was Kabusk getting coordinates and a photo from the photographer who had been there that provided the site match (Twitter), about 175 meters southwest of the investigated field impact.
Coordinates: 33º31'09.4” N, 36º21'17.3” E.  

The Spot is a basic visual match, with a strange darkness appearing on that roof and the one just west of it, sometime between Feb. 20 and August 23, 2013 (the dates of the Google Earth imagery closest to the incident date), matching the video in suggesting something ... (fire-related?) happened here, singeing the roofs, or some dark-colored mass of sarin fluid stained it? Anyway, I quickly verified this with select matching features noted (on Twitter, and see below) and, just before me, Michael Kobs verified it with other features. All relevant roof details, the bigger satellite dishes, the two little pillars to the west (green box), the raised walls on the building to the northwest, everything comes through where it should.

Now that we can place it, it's clear that this site was either not mapped, or mapped grossly wrong, by Human Rights Watch in their widely-cited map of 12 Zamalka-area impacts.
tally:
- dead sheep, and this, are verified as not mapped, or not mapped well.
- the two field impacts everyone agrees on: we agree there.
- the wall scene, w/some reservations, placed in that blue box - HRW may have it close by.
- garage door by Kobs, not verified by me, but likely right. HRW again isn't close, but two semi-close dots could be attempts based on flawed coordinates they got. The southern one of these is what Kabusk may have taken as a close mapping of his incorrect rooftop impact placement (a bit off in the other direction).

Direction of Fire

So, location being important to set it, we turn to the main point here - estimating the firing angle. In the end, the west wall faces almost due west, so the angle could have been set pretty well without the location, But where that angle of lines to is also important. 

Here's the scene again with some basic markers for reference. The red line marks the maximum width of missing wall. The impact center line is presumably around the center of this (red vertical) or maybe, my guess, on the line marked in purple. To me it seems low-centered on the wall; with impact likely on or near purple line extended up less than halfway, to a little below that odd jutting brick. Lots of upper wall gone could be just from the initial shock wave and then lack of support. The loss of so much lower wall is harder to explain without a low-centered impact. (that's a vertical angle clue, not as important here)

From there the rocket must have traveled towards the camera and left, besides down, into the floor at about the purple box, causing the wider damage marked in red. The purple area is a bit fudged between indicating the worst of the rebar damage (bars torn free and pushed far down) and the area at non-distorted prior floor level. 

Within the red area, the initial hit would be centered well left of middle, closer to the left-hand wall. Considering foreshortening in the video view, the center of damage is clearly a greater distance in from the west wall. Consider also: exterior photo from UN report, at right.

Mapping this on the satellite view: the raised part we see through in the videos is just at the northeast corner of the building, with the west half just open roof. The southeast corner is walled off, as the UN photo above shows, but seems to have no roof (it's full of sunlight). It seems the roofed corner is divided into two rooms, with the southern one impacted, and the entire width of its west wall knocked out. The small part remaining is at the southwest corner by the hole in the floor. So I map it like this:

This yields a suggested direction back to firing in a possible range of 299º to 323º, with a middling best estimate of 313º. This comes out right in the middle of a broad range Michael Kobs set, and puts it clearly in the set of northwest-origin rockets (see working draft of angles study here). Any direction closer to 350º, as we had been taking as the lone firing direction, is all but excluded in this case. This is an estimate; presumably all northwest firings converge on one spot, so likely the direction out is a bit more to the north, maybe like 318º. Considering the UN photo, I likely put the hole too far in; if so, the angle would point a bit more north, like 316 or 318º.

Note that while this is fairly readable, the OPCW investigators didn't get a direction reading here, at least according to the UN report that said only one of the 3 spots visited - the nearby field - gave a readable and "precise" direction of 285º. But that was grossly wrong, as anyone can easily verify (see part 1). This nearby spot would have been better to cite, with its direction a little more flexible. I can see the direction being as low as 299º, but as it happens, I doubt is could be quite what they need either. At the implied distance to regime missile base of about 10km, the difference in heading between here and the filed is nil. This would be 285 as well if it matched, and that's just a bit too low to be right.