Leaked OPCW documents: what they really show about the Douma investigation
Recently-leaked documents disprove claims by a whistleblower that the OPCW doctored a draft report of its investigation into a suspected chemical attack in Syria. They also cast doubt on claims that the views of staff were systematically ignored. They do, however, confirm a dispute involving one staff member who was eventually suspended and escorted out of the office.
In April 2018 more than 40 people were reportedly killed by a chemical attack in the Syrian city of Douma. Western powers blamed the Assad regime’s forces and launched punitive strikes against what were said to be the regime’s chemical weapons facilities.
After investigating for almost a year, the Organisation for the Prohibition of Chemical Weapons (OPCW) found “ reasonable grounds” for believing a toxic chemical had been used as a weapon in Douma and that the chemical was “likely” to have been molecular chlorine. It was also “possible”, the OPCW said, that the chemical had come from two yellow gas cylinders found at the scene.
Meanwhile, the Assad regime denies using chemical weapons in Douma or anywhere else and is supported in that denial by its chief ally, Russia, along with the regime’s defenders on social media. They reject the OPCW’s findings and claim evidence from its investigation was manipulated in order to reach a conclusion that would help justify the US-led airstrikes.
The regime’s defenders maintain there was no chemical attack in Douma. They say rebels faked the appearance of an attack, including the planting of dead bodies, in order to falsely incriminate the regime and provoke US-led airstrikes.
In October the information war over Douma took a new turn when a whistleblower surfaced. Identified only as “Alex”, he is said to have been a member of the investigating team who was responsible for collecting samples at the scene of the alleged attack.
Alex has made no public statements himself but his allegations have been reported by sympathetic third parties.
He has been quoted as saying that most of the Douma team felt the OPCW’s two reports on the incident — an Interim Report in July 2018 and the Final Report in March 2019 — were “scientifically impoverished, procedurally irregular and possibly fraudulent”.
He is said to have described “disquieting efforts to exclude some inspectors from the investigation whilst thwarting their attempts to raise legitimate concerns, highlight irregular practices or even to express their differing observations and assessments”.
According to an open letter from Alex’s supporters, “Suppression of internal debate and questioning within the investigation team appears to have been systematic.”
Based on this, the regime’s defenders conclude that the US was pulling strings behind the scenes, manipulating the reports towards a “pre-ordained” conclusion.
It’s clear there were differences of opinion among investigators over how to interpret the evidence from Douma and how to present it in the reports but the OPCW says there was nothing irregular about that — it was part of a normal (and healthy) discussion process. In a rare public comment on the investigation, director-general Fernando Arias said the OPCW encourages “serious and professional” internal debates so that “all views, analysis, information and opinions are considered”.
Although Alex was said to have documentary evidence supporting his claims, several weeks elapsed before any of it was made public. In the meantime his supporters established a narrative on social media in which his claims were treated as established fact.
Eventually, on 23 November, a single email was posted online by WikiLeaks and this was followed, on 14 December, by five more leaked documents. Now that they are available it’s possible to review the evidence in the light of Alex’s claims.
Of the six leaked documents, four relate to preparation of the interim report on Douma. There’s the first draft, a revised second draft, an email discussing the two drafts, and finally an exchange of emails from a day or two before the interim report was published.
THE FIRST DRAFT
By the time the first draft was leaked it had already acquired something of a legendary status on social media. Written by team members who had worked on the ground in Douma, it had allegedly been suppressed for not saying what the bosses wanted to hear, thus prompting managers to produce their own “redacted” and “doctored” version of it in a second draft.
Alex had also reportedly made some startling — but misleading — claims about the first draft’s content: that it described what happened in Douma as “a non chemical-related event” and concluded that “the signs and symptoms” of the alleged victims were “not consistent with poisoning from chlorine”.
Now that the document is online, it can be seen that the draft only said some of the signs and symptoms were not consistent with chlorine. Furthermore, it did not assert that there had been “a non chemical-related event”. What it actually said was that “a non chemical-related incident” was one of the possibilities the team had considered in connection with the reported deaths — adding that so far there was insufficient evidence to reach an authoritative conclusion.
As a result of Alex’s claims the draft report has been characterised on social media as revealing hidden truths about Douma — truths that the US and OPCW bosses would prefer to keep under wraps. But while it did raise a lot of questions about conflicting evidence much of its content was later included in the official published reports, including several key findings:
- It said laboratory tests found no evidence of a nerve agent such as sarin.
- It said chlorinated organic chemicals “which are not naturally present in the environment” had been detected in samples gathered from the scene.
- It said the two gas cylinders found in Douma “might have been the sources of the suspected chemical release”.
At first glance, though, it does look as if the claims of doctoring might be true. While the first draft had 116 pages the second draft had only 34 and the published version 26, so obviously a lot of material had been removed. But a closer look shows this difference is largely explained by the length of the first draft’s annexes. When annexes are stripped out, the first draft has 30 pages, the second draft has 16 and the published version has 11.
That still means there were substantial omissions and, thanks to the leaking of the second draft, we can now see what they were.
THE SECOND DRAFT
In claiming that the first draft was “redacted” or “doctored”, Alex implies that something irregular or perhaps sinister had been going on. But a comparison of its text with the second draft, and also with the published reports, shows this was not the case.
One thing to keep in mind is that the Interim Report was only intended as an update on the investigation’s progress, so there were questions about how much information to include and how much to hold over for the Final Report. In the event, substantial amounts of content that had been “redacted” for the purposes of the Interim Report resurfaced in the Final Report — either intact or modified in the light of further work. There are several examples of omissions that fall into this “held over” category:
Paragraphs 7.9–7.14 (in the first draft) are a technical discussion of chemical reactions in wood samples. Although omitted from the second draft they can be found in the published Final Report (paragraphs 8.9–8.15).
Paragraph 7.30 suggests a strong link between chlorine and one of the cylinders which was found on a bed. The relevant part says:
“The team noted that a slat of wood that was lying under the cylinder on the bed, part of which was taken as a sample, was quite wet and soggy. No chlorine gas was detected in the room. On analysis, this wood sample showed the highest content of chlorinated organic compounds of all the wood samples taken.”
While this is missing from the second draft, it appears in the Final Report with only a minor alteration (changing “quite wet and soggy” to “damp and softened”.
Paragraphs 7.42–7.67 give a narrative of events, based on interviews with witnesses. Although omitted from the second draft, this formed the basis for a narrative in the Final Report (paragraphs 8.43–8.69) which included extra material from five additional witnesses.
Two other omissions relate to areas of controversy or uncertainty where there were hopes the picture would become clearer as the investigation continued:
Paragraphs 7.31–7.34 discuss one of the cylinders and speculate about “a likely incoming trajectory”. The more speculative parts of this are omitted from the second draft along with a diagram of the cylinder and its location.
Paragraphs 7.68–7.91 are about the signs and symptoms of alleged victims and the possible causes of death. They discuss conflicting evidence and the types of chemicals that might or might not have been involved.
In both these cases the investigators had failed to reach a conclusion and recommended consulting outside experts. Both issues were addressed in the Final Report in the light of the consultations.
There’s a further deletion which is difficult to explain:
Paragraphs 7.23–7.26 describe how gas from a rooftop cylinder could have been dispersed downwards through the stairwell of the building and possibly down the building’s exterior. This is omitted from the second draft, together with its accompanying diagrams, and does not reappear in either of the published reports.
For the record, there are also deletions in the last section of the draft, headed “Conclusions”. It has been edited to take account of deletions elsewhere in the draft.
A LEAKED EMAIL
Not all the deletions appear to have been contentious but a leaked internal email from one of the investigators highlights three which it describes as “particularly worrisome”.
The email is dated 22 June 2018 — a couple of weeks before the interim report was released — and is addressed to Bob Fairweather, a Briton who at the time was the director-general’s chief of cabinet.
The sender of the email, who is unidentified but appears to have had a major role in preparing the first draft, focuses on the two areas where experts were being called upon in the hope of resolving uncertainties: over the signs and symptoms of alleged victims, and the positioning of the two cylinders.
The email appears to be arguing that the revised draft is misleading by omission. “Many of the facts and observations outlined in the full version are inextricably interconnected and, by selectively omitting certain ones, an unintended bias has been introduced into the report, undermining its credibility,” it says.
The email talks of “the inconsistency between the victims’ symptoms, as reported by witnesses and seen in video recordings” — which it says raised questions about the identity of any chemical involved. It also complains that discussion about “the placement of the cylinders at both locations as well as the relative damage caused to the impact points, compared to that caused to the cylinders” is “essentially absent” from the second draft.
However, the revised draft did not ignore these issues entirely. It noted “mixed reports” of what chemicals might have been used, but without commenting on them. (After further investigation the Final Report also failed to reach a firm conclusion. “It is not currently possible to precisely link the cause of the signs and symptoms to a specific chemical,” it said.)
The revised draft talked about the cylinders too but avoided discussion of how they came to be where they were found, saying more work was needed.
The third omission complained about in the email concerns a bibliography which it says had been attached as one of the annexes to “support and enhance the credibility of the work of the mission”. Its absence proved only temporary, however, as it reappeared in the Final Report with some additional citations.
The differences between the two drafts were not just in what was left out — there were textual changes too, and the email complains specifically about three of these:
● The management’s draft said that chlorine or another reactive chlorine-containing chemical was “likely” to have been released from two cylinders found at the scene. The email says that while the cylinders might have been the sources of the suspected chemical there was “insufficient evidence to affirm this”.
● A sentence in the management’s draft referred to “the high levels of various chlorinated organic derivatives” in environmental samples. “Describing the levels as ‘high’ likely overstates the extent of levels of chlorinated organic derivatives detected,” the email says.
● The management’s draft talked about a “reactive chlorine-containing chemical”. The email says this was technically wrong and a more accurate term would be “a chemical containing reactive chlorine”.
Based on what whistleblower Alex is reportedly saying, it might be assumed that no one paid any attention but a look at the Final Report shows all three points were eventually taken on board:
● The final report described the cylinders as a “possible” source of the chemical rather than a “likely” one.
● The final report avoided describing the levels of chemicals found in the samples as “high” but noted that some were higher than others — there were “raised” levels of some chemicals close to the suspicious gas cylinders.
● The final report adopted the email writer’s preferred phrase about substances “containing reactive chlorine”.
THE INTERIM REPORT
At some point between compilation of the two drafts and publication of the Interim Report, a decision was made to adopt a more minimalist approach. The apparent intention was to avoid detailed discussion of unresolved issues, partly in the hope that they could be resolved later but also to reduce the scope for public speculation or misinterpretation.
As a result, the published report was mainly about the mechanics of the investigation: how the Fact-Finding Mission (FFM) had been set up, the brief it had been given, security issues, and what the investigators did in Douma. Most of this was copied directly from the original (allegedly doctored) draft with only minor changes.
The Interim Report did, however, reveal two important findings: that laboratory tests showed no evidence of organophosphorus nerve agents such as sarin and that “various chlorinated organic chemicals” were found in samples from the two suspect sites.
One of the recently leaked documents is a collection of emails exchanged between members of the investigating team and Sami Barrek, the team’s Tunisian leader. Written one or two days before the Interim Report was published, they discuss last-minute changes to its text.
While it’s possible that “huge internal arguments” were taking place around this time, as the whistleblower claims, the emails don’t really show it. They show discussions among colleagues that — for the most part — are calm in tone and broadly constructive.
In an email to Barrek dated 4 July 2018, a team member proposes adding a few lines about the gas cylinders found in Douma, including a sentence saying: “Work is on-going to assess the relative damage to the cylinders and the roofs as well as the provenance of the cylinders and their possible trajectories.”
An hour later another team member chips in, suggesting “possible trajectories” should be changed to “provenance” or “how the cylinders arrived at their final resting places”. “Possible trajectories”, the emailer say, implies the cylinders were flying out of the sky — “which at this stage we are not necessarily concluding”.
Barrek appears to accept that and responds that he is also inserting a sentence at the end of the summary section saying: “The FFM still needs to clarify some of the details and to this end, the investigation remains on-going.”
There may have been further discussion of this which is not recorded in the emails, because the sentence actually published is slightly different. It says: “The FFM team needs to continue its work to draw final conclusions regarding the alleged incident and, to this end, the investigation is ongoing.”
One thing that emerges clearly from these emails is that investigators were worried about news media jumping to unwarranted conclusions or others misconstruing the findings to suit a political agenda. In the words of one of the emails: “It’s because the stakes are so high that we have a responsibility to guard against misrepresentation, by both sides.”
TRACES OF CHEMICALS
One area of contention in the emails is about the wording of a paragraph saying chlorinated organic chemicals (COCs) had been detected.
On the morning of 5 July — a day before the Interim Report was published — team leader Barrek sent out an email announcing he had made an executive decision.
“Good morning,” Barrek says in the email. “After reflecting on the second to last sentence in par 2.5, I decided to remove the detail we discussed two days ago about the concentration, especially that the next sentence covers the idea.”
The exact wording of the deleted sentence is unknown but it’s clear from the ensuing discussion that it said the COCs detected were in trace-level concentrations (i.e. very small). It’s also clear there was uncertainty among the investigators as to whether or not this mattered.
Responding to Barrek’s announcement, one of the investigators warns that leaving out the reference to concentrations could “allow some readers to arrive at a simplistic conclusion” — that the COCs were definite evidence of a chemical attack. “The ppb [parts per billion] concentrations is what qualifies the subsequent statement that we need to figure out whether these levels have any real relevance,” the email says.
An hour later Barrek replies: “I think we should keep the same level of details everywhere in the Interim Report. There are also other very important facts that we decided not to release in this Interim Report.”
Three minutes later another email from an investigator challenges Barrek’s decision: “Can we take it then that you are unilaterally deciding to remove this fact from the report against the recommendations of the team?”
Barrek responds: “I considered inputs from all team members; I have the support of most team members. I would like to remind you that I can take unilateral decisions, nonetheless I try to take into account inputs from everyone when possible.”
Half an hour later, another team member intervenes, attempting to mediate. In the email, the team member acknowledges that the risk of misinterpretation comes from two directions. Reporting the detection of COCs without mentioning levels could be “misinterpreted as a damning conclusion”. On the other hand, stating that the COCs were only at trace levels could lead to the results being “misrepresented by the other side of readers”. The email proposes a compromise wording: “The team is conducting further work and sample analysis to assess the significance of trace levels of chlorine-containing compounds.”
It ends by telling Barrek: “You may disagree with this approach, so please let us know your views. But I’d like to see if we can compromise … As we saw the other day, there are pressures and expectations from all sides. It’s a political wordsmithing situation, not purely technical.”
DEBATES ABOUT LEVELS
The issue of levels had already come up briefly in the 22 June email complaining about changes to the first draft of the Interim Report. The first draft said that materials sampled in Douma had been in contact with one or more substances containing reactive chlorine and that this was “based on the levels” of chlorinated organic derivatives which are not naturally present in the environment.
Why the first draft mentioned levels at all is puzzling because the obvious inference to be drawn from it is that they were relatively high — which may explain why, in the second draft, someone changed “levels” to “high levels”. The sender of the email (who was also a co-author of the first draft) objected to this change, saying:
“Describing the levels as ‘high’ likely overstates the extent of levels of chlorinated organic derivatives detected. They were, in most cases, present only in parts per billion range, as low as 1–2 ppb. which is essentially trace quantities.”
Even so, there’s nothing in the email to suggest that the complainer thought this invalidated the findings. An article on Bellingcat’s website about the chemistry of the samples suggests that trace levels are what would be expected in such a situation.
The trace levels have since figured prominently in reports of the claims made by whistleblower Alex. He has variously been reported as saying there were no extra COCs in the buildings where people allegedly died than in the normal streets elsewhere, and that they were much lower than what would be expected in environmental samples.
So far, though, Alex has not provided figures to support these claims and the FFM does not appear to have collected samples from other parts of Douma for comparison.
THE HENDERSON AFFAIR
In line with its recommendation in the Interim Report, the FFM commissioned three independent experts to investigate the two cylinders in more detail. Photos suggested they had both caused holes by hitting concrete roofs, with one of them ending up on a bed in the room below. Using computer modelling and other advanced techniques, the experts concluded that the damage at the scene could indeed have been caused by the cylinders crashing into the concrete roofs. Their findings were described in the Final Report.
However, in May this year — a few weeks after the report was published — a 15-page document surfaced on the internet. Headed “Engineering assessment of two cylinders observed at the Douma incident”, it was an internal report by Ian Henderson, a South African who had been one of the first inspection team leaders recruited by the OPCW after its formation in 1997.
With assistance from two universities (which the OPCW paid for) Henderson had carried out his own computer modelling — and reached a different conclusion from the FFM’s experts. His report argued that the Douma cylinders were more likely to have been “manually placed” in the spot where they were found than dropped from the air. This caused a stir among the chemical weapon deniers because it could be interpreted as support for their claim that rebels had faked a chemical attack.
The leaking of Henderson’s report immediately raised questions as to why no mention of his findings had been included in the Final Report.
After some hesitation the OPCW said his report had been considered along with all the other evidence but that it had caused a problem relating to the FFM’s mandate. The official role of the FFM is to establish facts about suspected chemical attacks but not to attribute blame. In the OPCW’s view, Henderson’s idea of manually-placed cylinders “pointed at possible attribution” and thus fell outside the mandate. It added, though, that Henderson had agreed to pass his document to the newly-formed Investigation and Identification Team which does have the power to attribute blame and would therefore be able to consider it.
Henderson appears to have played a central role in raising questions about the Douma investigation, especially regarding the OPCW’s conduct of it, and so far he is the only one to have been identified by name. During the debate about trace-level COCs, he was reportedly one of four investigators who initially persuaded team leader Barrek to mention trace levels in the Interim Report — though Barrek rescinded the decision two days later in his 5 July email.
Henderson’s relations with his employer eventually became so bad that in May this year he was suspended and escorted out of the office. Two documents leaked among the latest batch have shed some light on his side of the story.
The first document is a memo dated 14 March 2019 — two weeks after the Final Report was published. It’s addressed to director-general Fernando Arias and although the sender’s name is blanked out in the leaked copy the text leaves no doubt that the sender is Henderson.
Henderson says his first purpose in writing is to inform Arias that “there are about 20 inspectors who have expressed their concern over the current situation”. The “situation”, he says, is that the Final Report “does not reflect the views of all the team members that deployed to Douma”.
One of his complaints is that the authors of the final report, with one exception, were members of the investigating team who had not been working on the ground in Douma.
According to Henderson, when the team members deployed to Douma finished their work in Syria, the consensus among them was that there were “indications of serious inconsistencies in findings” but “after the exclusion of all team members other than a small cadre of members who had deployed (and deployed again in October 2018) to Country X, the conclusion appears to have turned completely in the opposite direction.”
The memo then moves on to Henderson’s own grievances …
“Upon my return from the FFM mission deployment to Douma (and my subsequent manning of the Command Posts in Syria for an additional 5 weeks), I was asssigned the task of analysis and assessment of the ballistics of the two cylinders. I undertook this in the understanding that I was clearly the most qualified team member …
“In subsequent weeks I found that I was being excluded from the work, for reasons not made clear … I was also assigned to other missions. However I made clear that I would complete the work and submit my report to the FFM.”
He goes on to describe how he “engaged engineering expertise, to get access to sophisticated engineering computational tools” for his report. Henderson says the work was approved by the Director of Inspectorate, and there’s no reason to doubt that. Judging by his memo it sounds as if others later tried to stop him from continuing but he was not to be discouraged and insisted on finishing the job.
Henderson’s official status within the investigation is a matter of some dispute. When his “engineering assessment” was leaked the OPCW’s press office initially said he had never been a member of the FFM, though others insisted that he was.
While he may not have been an official member, he was certainly involved. The OPCW later described him as a liaison officer at its Command Post office in Damascus. “As such, and as is customary with all deployments in Syria, he was tasked with temporarily assisting the FFM with information collection at some sites in Douma,” it said.
An email among the latest batch of leaked documents gives more details of Henderson’s activities with the FFM. Dated 20 May 2019, it comes from an unidentified supporter and is addressed to Veronika Stromsíková. the OPCW’s Director of Strategy and Policy:
“A member of the FFM team has been suspended from his post and escorted from the OPCW building in a less than dignified manner … Ian Henderson’s personal and professional integrity have taken a knock in the most pubic of fora, the internet. A falsehood issued by the OPCW, that Ian did not form part of the Douma FFM team, has been pivotal in discrediting him and his work.
“The denial is patently untrue. Ian Henderson WAS part of the FFM … He did join towards the end of the mission but in the few days he was there he deployed to Douma on three occasions: to a suspected chemical weapons production facility, to the hospital which was central to the alleged incident, and to Location 4, one of the sites where an alleged chemical attack took place.
“He subsequently deployed to a fourth location to re-examine the cylinders and apply seals. He was, in fact, the only engineer in the team and consequently the only one with the skills to do a proper engineering assessment.”
This still leaves the mystery of the OPCW ended up with two technical assessments of the cylinders being conducted at the same time — one by Henderson and the other by the FFM’s experts.