Rocket Man: Copied Impact Site Mix-Up

What the "Rocket Man" Wants us to See in Ghouta: Copied Impact Site Mix-Up

August 9, 2021

rough, incomplete

While our currently controversial Ghouta report(s) co-author Chris Kabusk is quite gifted at at visual geolocation, he's not perfect at it. One error he made in or by Feb. 2014 finally became clear only in 2017, raising a new mystery with a fairly obvious answer, but one that helps to get agreement on. Only in Feb. 2020 did I help Chris confirm the UN-OPCW site 5 where the north-facing balcony of a vacant apartment was impacted by a rocket we don't see (except likely bent on the street below).

He had already accurately linked this spot to a Zamalka street scene of fully-protected OPCW inspectors. They're looking down on collected samples, so it's following a visit to some CW-relevant site nearby. At right is my color-coded update of his matching - photo taken from the apex of the blue angle of view. 

The opposition LCC and HRW had both mapped an impact just a little behind them here, suggesting it was that site the inspectors were just at. It's likely imprecise as all of those mapped spots,  so LCC2 might refer to any building in the block on the left (5 buildings seen here), or somewhere nearby. A red triangle marks the corner of the building decided by this graphic, as used in my June 2021 report. Do note original draft (to be updated) messed up the coordinates for this spot. The correct ones for the spot shown are 33.5216642° N, 36.351963° E.

There were TWO impacted buildings the OPCW visited in Eastern Ghouta, out of three impacts total: sites 3 and 5 as they called them - besides the field impact site 4. Site 5 is barely mentioned, maybe not even named in the official report (PDF), maybe since nothing but small pieces of the munition were found here. But they specify five impacts were looked at, note a visit to a final lone site with a balcony on August 29, and they were filmed there that day. Sites 3 and 4 were visited simultaneously the day before, on the 28th, both also filmed (not, I think, to the inspectors' liking, but usefully for us - thanks, Islamist morons).

There were also two street scenes with the inspectors that could be located, maybe correlating with sites 3 and 5. But they also visited a clinic, matching best with the other street scene (right) in which they are not geared up with gas masks. So just the one street scene (again, the one above) should relate to EITHER site 3 or 5. 

Reference: all LCC, HRW, and collective impact mappings compared:

Deciding on site 5 as we did involves deduction, and the exact building may be wrong even in the graphic above. I may have misread a brief view between buildings - not really clear which building then - perhaps the one just NE, with "impact" label resting on its edge - for future review. Here are the views of the spot I need to re-consider: left 3 frames from a video I've seen, right before the street scene in question (thus likely connected, but uncertain) - right: three crops from a video I haven't seen, via Bellingcat (see below) - colored marks clarify they all show the same spot moments apart - Time if site 5: about 1:50-2PM; last samples on 29 Aug. were taken 13:46 local time (report, p. 26). If site 3 (possible), similar time anyway.

So guesswork is still involved in that, but this area is much better narrowed as the home of impact site 5 now that the other building impact - site 3 - has been undeniably placed well to the south in Ain Tarma. Chris Kabusk did that using site photos, back in August, 2017, and we worked together confirming several totally matching features seen from the other view to the west. Update: Chris did NOT use coordinates in that 2017 search, as I wrote. Apparently I was confused between several questions and several answers. He says he used visuals alone for that, in the other photo from the same roof the photographer shared with him (photo and site match below). 

Impact site 3 coordinates: 33.519130°, 36.354841° - or as shown above, or similar will point to that same rooftop. Those last digits are very fine-tuned. Actual first impact (from the northwest) was roughly in the middle, over open roof on the west side, into a N-S wall of this partly-roofed 1/2 of a 6th floor.  

But before site 3 was set there, we had video & photos but no description for site 5, and for site 3 we had unplaced visuals, and the UN report describing it: a rocket "penetrated a cinderblock wall and a rebar containing concrete floor before coming to rest in a room below." That cinderblock wall was "on a roof of a 5-story building" that was visited August 28 by one team, while another team investigated impact site #4 in a "nearby open field," - so sites 3 and 4 are close by. 

Otherwise, it seems it might have been atop on of those other 5-6-story buildings a ways north, by the plotted LCC2/HRW2. In fact, it's been taken widely as fact that it was right here 

How it Was Done Wrong

In a bit of deduction that proved wrong, Chris had initially decided that street scene was just outside the roof impact site 3 (the only site described as at a 5-story building) and possible fit - and NOT what it eventually proved to be - the apartment balcony impact in one of those 5-6 story buildings. It had a small structure in the middle of the roof, as other buildings here have, perhaps enough to that one smashed into at impact site 3. 

Further, there was that mapped claim of an impact right about there. In contrast, the spot it wound up being ... was never mapped at all by LCC or HRW. So maybe expecting such a shortcut wound up limiting Chris' view. Indeed, as he said in 2017 "Yeah, I took a shortcut because of HRW's identified spots and also it was Eliot who told me that the rocket landed in Zamalka." (the identified spot is in Ain Tarma, like the "impact site 4" it was said to be near) 

Anyway, these are all fairly simple, reasonable mistakes to make. He had decide this early on, with one of his earlier tweets, Feb. 20, 2014 showing a roof very near eventual site 5 pinned for "rocket 3 Volcano" - it's in the middle image, sideways and distorted, but same as shown more clearly on 8/30/14 (and below) with 3D model - (modeled as 4-story - mislabeled UMLACA 5 not 3, mix-up noted - small labels: LCC4 (but it's #2 in the only numbered LCC map I have - see below. All these ref codes can get hopelessly confused. But otherwise...). This is the same spot looked at above - in fact the same square-ish building just NE of the current site 5 impact label - coincidentally, this might be the building with impact site 5. But site 3 was never here.

So this site was Chris' guess from at least Feb. 2014. However, the rooftop structure here probably isn't big enough. It's not very near to site 4 as described. The view to the west didn't match - buildings almost touching site 3's roof were clear enough even in 2013 views, while this has an open space to the west. Still, it seemed good enough to Chris than, and to most observers forever. He may have noticed any of these issues before, but at least when he saw the right spot emerge, he was capable of following those better clues - this time to a 100% certainty. But not everyone's mind is so capable, or willing to adapt to the 3D reality of events, whatever it winds up being. 

How that Error was Mindlessly Copied

Bellingcat founder Eliot Higgins variously cited, copied, verified by, or otherwise did the same as Chris on several points from at at least Jan. 30, 2014 and up to at least September, but in a different overall spirit. He did give public credit for some of this (even "big thanks"), but not all. And while I'm not sure it was Chris' exact error about impact site 3 that caused it, this is the same exact site 3 vs. site 5 mix-up Higgins and/or his team used and is still stuck on, down to the exact building. 

-- to fill-in -- it seems Chris was on positive terms with Higgins early on, even being invited to a Storyful newsroom events for journalists to share findings, network, etc. on Ghouta and other subjects. (Chris also retweeted on a Higgins presentation 2/19/14). As linked above, this site was squished into a Feb. 20 image, so available from then, and it was shown more refined in late August. On the anniversary, 8/21, he tweeted  a mapping update including "UMLACA 5 [suspected videos] UN inspectors." (He meant UMLACA 3 or actually Volcano 3, not 5. And he thought it was impact site 3 at UMLACA 3 but it was impact site 5. Just to be clear, lol.) So this finding was around all year, and no one else had provided a location guess.

Bellingcat's "Locating the Rockets Used During the August 21st Sarin Attacks in Damascus", (August 10, 2014) somehow only relates 3 of the five impacts then mapped, skipping the pivotal impact site 4 (so they didn't have to go on the record with a visual reading to match or clash with the one reported by the UN-OPCW mission). The one involving dead livestock was also skipped. 

But this explains the re-dubbed "rocket 1" (field on my map above), "rocket 2" (wall 2), and it closes with "rocket 3" as it was then called. This includes a video of the site with inspectors inside, looking at rocket remains that fell down next to a coffee table, with the midday sun beaming down through the hole it made in the roof/floor above. This is impact site 3, beneath the pierced roof (below: video paused outside the stuck door).  

Higgins goes on to explain "the key" with rocket 3 was "establishing the location of the building the video was filmed in." Well, he never got that key then. 

He goes on there to note as I have here "Only one group of videos filmed outside buildings matched" ... with ONE of two building impacts, by deduction. It's the scene filmed - as Chris had found - next to LCC2/HRW2. Next, "we were able to make out the following three individuals present both at the inspection site and outside one building." (One is a leader of the armed militant minders Liwa al-Islam provided who guarded these useful infidels around to investigate - as it turns out - Liwa al-Islam's own crime). What he and Bellingcat apparently didn't consider: maybe they were seen outside a different building in those separate scenes. 

But "Based on having established the location" - of the street scene not clearly linked to any other exterior scene, really, nor to either interior scene ... "it appears the inspectors are coming out of the building to the north of the building marked by the VDC and LCC..." That's in a video not linked, but using the stills shown above. It appeared to ME like the building to the SW marked in red above, but now I'm not so sure (future review). He might just be right on the building they're leaving, but it shouldn't really appear like any particular one without good reason, which Higgins doesn't seem to have. I think his reason was Chris having chosen that building, and he might have had good reason.

And "so it seems likely this was the building hit by the attack." One of the 5-6-story buildings here was hit, yes, but not with THAT rocket. That coffee table has been clearly placed way to the south. 

We fixed this error, but Bellingcat never did, and they're the main ones everyone else mindlessly follows. For example, in 2020 the Syrian Center for Media and Freedom of Expression (SCM), Open Society Justice Initiative (OSJI), and Syrian Archive brought lawsuits against Syria in Germany, France, and later Sweden so hopefully "Syrian officials can be arrested” or until then, new economic sanctions can be imposed, etc.. OSJI lawyer Steve Kostas heralded "the most detailed investigations to date into these attacks," which they had assembled to bolster these acts of hostility. Their factual basis was spelled out, in varying detail, in a public summary of the evidence released by SCM/OSJI/Syrian Archive. 

https://www.justiceinitiative.org/uploads/4bc2c0f7-81ad-48c5-8166-281ab35301ab/sarin-complaint_ghouta-public-summary_10062020.pdf

Regarding the rocket impact analysis, as my report had recently summarized:

"The public summary’s analysis to “verify the locations” of Volcano rocket impacts seems to just copy others’ open-source work, including a widespread mix-up between UN-OPCW sites 3 and 5. As such they they “verified” just four correct but pre-verified locations, “verified” one to the  wrong spot, and just “identified” one impact (it belongs at that wrongly taken spot, and was left unplaced). The report by three well-funded groups, with three staffs, explained that to achieve so little required reference to “hundreds of videos” and “working with the Human Rights Center Investigations Lab at UC Berkeley School of Law” as well. 

(By those definitions, the present report “identifies” one impact, “verifies” eight to their correct locations, and adds serious trajectory analysis. ..."

The public summary gives coordinates for the five impacts they considered, along with a description of each. One entry says: "Impact site C: a rocket pierced through the roof of a five-story building and landed in the room below, visited by the UN Mission on 28 August 2013,31 at coordinates 33.521641, 36.352425." They describe impact site 3, but give the coordinates for impact site 5, absent rocket penetrated the north-facing wall of a vacant apartment. Citation 32: "Coordinates by Bellingcat and Human Rights Center Investigations Lab at UC Berkeley School of Law. Rocket described by Bellingcat as “Rocket 3” in Bellingcat Rocket Analysis." 

This Human Rights Center Investigations Lab at UC Berkeley School of Law that helped with the location work - along with Bellingcat and doing it just the same as they did - might appear like added, independent confirmation. They show 5 placed impacts with no explanation how it was done past "By comparing videos and images of the attack with satellite imagery of Eastern Ghouta, the team confirmed the location of four previously identified impact sites and located an additional impact site. Each can be seen here." it's done just like Bellingact, including their mix-up of UN-OPCW sites 3 and 5. - that's their site C - they mean that's the rooftop impact, based on a bad guess they re-made or just copied. The public summary was wiser to leave that one just "identified." 

https://storymaps.arcgis.com/stories/56c19f1dbcbb4054b524cacc5f6a9fa5

But really it seems more like the Berkeley team just copy-launded Bellingcat's prior findings. Being a 2020-era view, it is interesting to see if it was improved since 2014. Nope. As named here and in that public summary, it's "impact site C" that's set quite a bit north of where it actually was. It's wrong just like Bellingcat was, just like Chris was at first, but without the subsequent learning he and 'his new team' had added - THREE YEAR BEFORE this blunder here. 

In Conclusion

Ironically, if Chris Kabusk's work led anyone astray with this error, it's the same people (loosely speaking) who now attack his credibility as likely distorting or disqualifying our work. Rather an unrelated imperfection in his early analysis - plus their own laziness caused Bellingcat to leave this error intact. They never formed the doubts Chris did about that site match. They never followed up on the 2017 photo clues. They never even copied our work along those lines between 2017 and 2020. 

And so the same easily-refuted mistake was passed on lazily to the kids at Berkeley, the SCM, OSJI, Syrian Archive and that pompous lawsuit. It's not some stray error either - the public summary admits the rockets could only fly ~2km, but also admits to including witnesses for them being launched from Mt. Qasioun, 10km distant. They have false claims, impossible claims, and maybe even a few true ones, all agreeing the "Assad regime" must be blamed and punished yet again.

Higgins for Bellingcat had boasted, in that 2014 article, of their careful correlation work on the clincher rocket 3, which was "more of a challenge" than the others. Of course he loves a challenge, like a complex video game. He and/or "we" met it as shown, and says "this demonstrates how a variety of information can be used to establish important facts that might otherwise be impossible to establish." 

No. It shows how a decent try can also establish non-facts like this. We can all run into this. This very blog, hard to believe (lol) is still full of wrong calls I know of and only have time to update sporadically or in the next version. The better minds notice it and correct - ideally more uniformly than I do. But other minds just stick to their beliefs, rational or not. And Eliot Higgins seems to really believe in Chris Kabusk's work, like to a fault. But overall, that's for good reason. 

As he continued on how this all pays off (same 2014 article): "As I demonstrate in “Identifying Government Positions During The August 21st Sarin Attacks” this information then becomes part of a broader understanding of the attacks, giving us a clearer understanding of where the attacks could have originated from." As he finds over and over with various faulty readings, that can include a number of government-controlled areas. Next post.