3.1 Witnesses

The Final Report provides no information about how witnesses were selected. Earlier FFM reports that interviewed witnesses in Turkey mentioned organizations that had selected these witnesses. For the report on the alleged chemical attack in Khan Shaykhun in April 2017 these organizations included: a Belgian non-profit company named Same Justice / Chemical Violations Documentation Centre Syria that has never complied with the legal requirement to file accounts;, the UK government-funded White Helmets; and the Syrian Institute for Justice coordinated by Jamal Juneid who had been in charge of external relations for the “Free Governorate of Aleppo” till 2016.

The Final Report records that between 14-22 October the FFM redeployed to conduct interviews: no information is given about why these additional interviews were conducted, or where and with whom they took place. As there is no record of another deployment of the FFM to Syria after 2 May 2018, we can infer that they were conducted outside Syria with people affiliated to the opposition, and (by comparing the interim and final reports), that they included five individuals not interviewed during the first deployment to Turkey in April and May. The FFM does not reveal whether any of the 21 individuals interviewed during the first deployment to Turkey were re-interviewed in October, or whether their stories had changed between May and October.

The information obtained from all interviews is summarized in 26 paragraphs, without any distinction between information obtained from those who had remained in Damascus and those who had relocated to Idlib under the evacuation agreement (who would have been opposition supporters), and without any effort to establish whether the presence of each witness at the scene was corroborated by other evidence. For instance this paragraph reports contradictory testimonies without any attempt to establish which of these accounts is compatible with the images recorded at the hospital:

Some witnesses stated that many people died in the hospital on 7 April as result of the heavy shelling and/or suffocation due to inhalation of smoke and dust. As many as 50 bodies were lying on the floor of the emergency department awaiting burial. Others stated that there were no fatalities in Douma Hospital on 7 April and that no bodies were brought to the hospital that day.

3.2 Environmental and biomedical samples

The report states that:

All the environmental samples from Douma were collected by the FFM team on Syrian territory in the presence of representatives of the Syrian Arab Republic.

The Annexes record that the FFM also received environmental samples from witnesses on 18, 21 and 22 April. If the statement above is correct, these samples can only have been received outside Syrian territory. The Annex lists four items received on 21 and 22 April: pieces of timber, scarf from basement, stuffed animal from basement, piece of clothes from victim. It is not clear what evidential value these items could have had, with no chain of custody, when samples from the sites of the alleged attacks could be taken directly.

The samples were analysed in two batches:

The results of analysis of the prioritised samples submitted to the designated laboratories were received by the FFM team on 22 May 2018 and 8 February 2019.

No explanation is given for why “prioritised samples” collected in May 2018 should have taken more than eight months to analyse.

3.3 Images

The Annexes list 206 media files without any key to content, upload times or metadata, of which 113 were reported to contain metadata. The report stated that only files that contained metadata were evaluated:

Understanding that many of the same sources of information are available online, material reviewed by the FFM was provided to the FFM by casualties, witnesses and medical staff. Only digital information that contained metadata was evaluated for the purposes of this report.

A description of images from the hospital dousing scene appears to contradict this by stating that only “some videos” contained metadata:

Some videos contained metadata and were recorded approximately three hours after the reported time of incident.

A natural reading of this sentence, interpolated into a description of the scene, is that this description is based on videos with and without metadata. There are more contradictions between the main body of the report and the Annexes. The report states that:

Videos taken inside Location 2 were recorded between 13 and 16 hours after the reported time of the incident, based on retrieved metadata (Annex 11)

This is contradicted by a chart in the Annex showing that the first images of “decedents” were timed to shortly after 10 pm on 7 April, consistent with other reports that the timestamp of the first video from Location 2 was 10:06 pm. This is three hours, not “between 13 and 16 hours” after the time of 7 pm that the FFM gives as “reported time of the incident”.

A note in the Annex states that:

From an examination of the metadata, the videos and photos provided by witnesses in relation to Locations 1, 2, and 4 were created at the reported time of the alleged incident.

This contradicts the statements in the main body of the report that videos of the hospital dousing scene “were recorded approximately three hours after the reported time of incident” and that videos at Location 2 were recorded “between 13 and 16 hours after the reported time of the incident”.

Thus although the report has excluded files without metadata (though even this is ambiguous in the description of the hospital dousing scene), it has included files from Location 1 and Location 2 for which the timestamps do not match the reported times of the incidents, and it makes what appears to be an unequivocally false statement that the timestamps match the reported times. The FFM was well aware that timestamps cannot be relied on, having concluded that the content of four files timestamped 12 June 2015 was nevertheless “related to the incident on 7 April 2018 in Douma”,

Timestamps of media files may be incorrect for various reasons: the device may have had its clock set incorrectly, the file may have been edited and saved, or the metadata may have been deliberately altered. A serious analysis of this material would have included all media files with or without timestamps, and would have tabulated for each file the timestamp if any, the earliest record of an uploaded image (which gives a latest bound on the time of recording), and the evidence of timing contained in the images. This visual evidence would include visible clocks/watches, calculation of solar time from shadows or twilight stage, and arranging images of the same scene in sequence based on continuity.

3.4 Failure to proceed with exhumations

The Interim Report of the FFM, released on 6 July 2018, stated that the intention to exhume bodies from mass graves was “communicated to the Syrian Arab Republic in Note Verbale NV/ODG/214827/18 and preliminary preparations were undertaken by the Secretariat for this eventuality.” The Final Report gives the date of this Note Verbale as 26 April, and states that:

The Syrian Arab Republic replied in Note Verbale No. 45 on 4 May 2018 and enumerated the conditions to be met in order to conduct the exhumation. With due consideration of the time elapsed since the alleged incident, the possibility was eventually not explored any further.

If the FFM had decided before July not to proceed, we might have expected this to be included in the Interim Report released on 6 July. If the FFM had not decided at this time whether to proceed, we might have expected some explanation for the delay in reaching this decision. Although the time elapsed might have limited the value of autopsies or tissue sampling for determining the cause of death, it would still have been possible to obtain DNA samples, which might have allowed victims to be identified through matching with living relatives and with each other. Other identifying information might have been obtained from clothing, items in pockets or X-rays. Establishing the identity of the victims would have been critical in determining whether those who came forward to give interviews reporting that their relatives had died at Location 2 were telling the truth.

3.5 Consultations with experts

The delay from October to December is attributed to consultations with unidentified toxicologists and engineering experts, with “reception of engineering studies” in December. There is no explanation for why, if the FFM had considered it necessary to obtain expert opinions on the possible trajectories of the cylinders found at Locations 2 and 4, they did not request on-site examinations while the FFM was deployed in Damascus in April/May, rather than six months later when inspection of the sites with cylinders in position was no longer possible. No information is given on how the experts were selected, how possible conflicts of interest were managed, or whether the opinions quoted represented the consensus of all experts.

The experts were asked to provide assements on “the trajectory and damage to the cylinders found at Locations 2 and 4”. For Location 2 “the expert provided reports and numerical simulations on the impact of steel cylinders against reinforced concrete slabs”. For Location 4: “The team consulted experts in mechanical engineering, ballistics and metallurgy to provide qualified, competent assessments of the cylinder trajectory. A natural reading of this is that these experts were not asked to assess whether the holes in the roof and the positions of the cylinders could be accounted for by anything other than being dropped from the sky. No technical details of the simulations or calculations are given: the figures lack explanatory captions to describe what they show. Figures labelled”Computer modulation of the aperture and cylinder impact” and “Diagram demonstrating the possible movement of the cylinder at low speed” appear to model the deceleration of the cylinder at Location 4 during the first few milliseconds after impact. These models appear to be based on the cylinder striking the roof at 60 m/s, or on a fall from 150 metres. No basis for these modelling assumptions is given; a Russian expert has pointed out (at 25’30″ in the video) that Syrian helicopters over a defended area would fly at an altitude of at least 1000 metres, and that a cylinder dropped from such a height would strike the roof at far higher velocity. To account for the size of the hole in the roof, the FFM postulates that the cylinder struck the roof with its long axis parallel to the ground, broke through and bounced sideways off the floor of the room on to the bed:

The studies further indicated that, after passing through the ceiling and impacting the floor at lower speed, the cylinder continued an altered trajectory, until reaching the position in which it was found.

No visual evidence of any impact on the floor is shown, and no explanation is given as to how a cylinder falling vertically on a flat floor would be deflected sideways. The cylinder is in a harness with fins, and the cylinder valve is still present. The FFM does not explain how the contents of this cylinder were discharged if the valve was intact and the cylinder was not ruptured. An expert opinion provided to us and to others noted that if the cylinder had broken through the roof with its axis parallel to the ground, the valve and the fins would have been sheared off.

4.1 Staging of hospital dousing scene before alleged attack

One of the videos uploaded from Douma was a scene in a hospital where children were shown being doused with water, supposedly to decontaminate them after a chemical attack. The children in the video showed no signs of exposure to a toxic chemical. A briefing by the Russian Defence Ministry on 13 April 2018 showed a clip of an interview with two medical staff who were seen in this video, and reported their account of the scene:

those who had been brought to the hospital had no injuries caused by chemical agents. When civilians were receiving first aid, some unidentified people rushed in the hospital. Some of them had video cameras. These people started shouting, spreading panic, and dousing other people with water. They shouted that all people in the hospital were victims of chemical weapons use. Patients and their relatives started dousing each other with water. After the action was filmed, the unknown persons escaped immediately.

One of the children shown in this video was identified as 11-year old Hassan Diab, who with his father told his story in an interview with Rossiya 24 on 18 April 2018.

We were in the basement. Mama told me that we have nothing to eat today, we will eat tomorrow. We heard people shouting in the street – go to the hospital. And as soon as I walked in, they grabbed me and started pouring water on me. Here’s me in the video.

His father added that “Militants gave them dates, biscuits and rice for participating in this film and released everyone to their homes.”.

The OPCW Director-General stated on 16 April that the Syrian authorities had offered 22 witnesses who could be brought to Damascus. The Interim Report stated that 13 interviews with witnesses had been conducted in Damascus, but gave no further details : “Analysis of the testimonies is ongoing”. The Russian representative to OPCW reported that six of the 17 witnesses brought to The Hague by Russia for a press conference in 26 April had already been interviewed by OPCW:

Six of the Douma witnesses brought to The Hague have already been interviewed by the OPCW technical experts, Russia’s permanent representative to the OPCW, Aleksandr Shulgin, said. “The others were ready too, but the experts are sticking to their own guidelines. They’ve picked six people, talked to them, and said they were ‘completely satisfied’ with their account and did not have any further questions.”

The Final Report appears either to deny that these witnesses had been formally interviewed by the FFM before they were brought to The Hague, or to have decided that their subsequent appearance at The Hague was a sufficient reason to exclude the testimony they had already given in formal interviews:

The statements from alleged witnesses who came to The Hague (presented in some media, see Annex 2, second bullet point) were dealt with by the FFM as other open-source video material.

From other paragraphs however it appears that the FFM has included interviews with some of the medical staff who testified at The Hague:

A number of the interviewed medical staff who were purportedly present in the emergency department on 7 April emphasised that the presentation of the casualties was not consistent with that expected from a chemical attack.

The report conveys doubt about whether these witnesses were present in the hospital by using the words “alleged” and “purportedly”, without reporting whether their presence was corroborated by videos or other evidence. The FFM reports that the hospital dousing scene was timed by witnesses to “shortly after” 7 pm, and quotes only one eyewitness account of this scene:

A witness reported that he was asked at the emergency department to help hospital staff to wash casualties and, while performing this task, a man who was not from the hospital entered, shouting “Chemical! Chemical!” and panic ensued. Bystanders then began undressing and washing people and proceeded to give inappropriate treatment.

The presence of “bystanders” and cameras is unexplained.

The report describes one or more videos showing the hospital dousing scene:

A video taken at the medical treatment facility depicts approximately 20 people (males, females, adults and children) being treated in what appears to be a temporary facility. Some videos contained metadata and were recorded approximately three hours after the reported time of incident. Simple decontamination procedures (washing with water) are carried out on a number of adults and two to three children roughly three to five years of age. Any distress displayed is noted to be mild. There are three young children of approximately 12 to 18 months of age (one male, one female and one of unidentified gender), each of whom is displaying objective signs of respiratory distress manifesting as laboured breathing and accessory muscle use. None appear to be cyanotic. One (male) child is intubated and seen to be receiving manual ventilation and later mechanical ventilation. The other (unidentified gender) child is seated partially upright with an adult and is being treated with a simple oxygen mask. The third (female) child is unresponsive with accessory muscle use, sluggish pupils and miosis estimated to be approximately three millimetres in diameter. She displays no objective signs of hypoxia. Multiple children are seen being treated with an unknown medication via a metered dose inhaler or small volume nebuliser. The adults and remaining children being treated in the video show signs of mild respiratory distress and coughing. No critically ill patients are seen aside from those paediatric patients previously described. There are no visible signs of external trauma.

This description is difficult to follow: the paragraph starts out by describing “A video taken at the medical facility …”, then interpolates “Some videos contained metadata and were recorded approximately three hours after the reported time of incident” before continuing with a description of what may or may not be the single video referred to in the first sentence. The description of the scene matches that in the uploaded video in which Hassan Diab was seen being doused with water except that the only individuals described as being washed with water are “a number of adults and two to three children roughly three to five years of age”. There is no mention of the older children, who included Hassan Diab.

The time of the alleged chlorine attack at Location 2 had been given as 8:22 pm by the White Helmets in a tweet at 5:42 am (local time) on 8 April, and as 9 pm by Dr Zaher Sahloul of the Syrian-American Medical Society. The video showing the cylinder apparently frosted by autorefrigeration was timed at 10:06 pm.

If it were established that the hospital dousing scene at about 7 pm was staged and that it was planned before the incident at Location 2, this would be strong evidence that the incident at Location 2 was staged. For clarity, we spell this argument out. Under the hypothesis of a real attack at Location 2, it would be a highly improbable coincidence for this attack to occur shortly after a staged casualty scene at the hospital. Under the hypothesis of a staged attack at Location 2, it would be highly probable that a casualty scene at the hospital would be staged at about the same time, possibly beforehand.

The FFM has distorted the evidence on several points:

• Hassan Diab’s testimony that he was lured to the hospital for the dousing scene has been excluded, and the description of the video recordings omits his role.
• Instead the report presents the statement of a single unnamed witness in such a way as to suggest that the dousing scene was a spur-of-the-moment panic.
• Without any explanation of why the earlier reports giving the time of the incident at Location 2 as after 8 pm were wrong, the FFM has brought forward the time of the incident at Location 2 to 7 pm, even though the first images from this site showing a cylinder still frosted were timestamped after 10 pm.

The effect of these distortions is to obscure the evidence that the hospital scene was staged and the implications for staging of the alleged incident at Location 2. This staging would most likely have been carried out by bringing to Location 2 the bodies of victims who had died somewhere else. This would imply either a managed massacre, or possibly an opportunistic staging, recycling bodies of civilians who had been asphyxiated elsewhere while sheltering from bombardment.

4.2 Fire lit before cylinder had discharged its contents

The Interim Report referred only obliquely to the visual evidence that a fire had been lit in the room below the cylinder at Location 2, listing “wipes taken from the burnt wall” among the samples collected. The Final Report makes clear that a fire was set, and reports an explanation given by a witness:

The FFM team noted the blackening of the ceiling and the rim of the aperture from the room immediately below the point of impact (see photo above). It also noted the blackened sooty walls in the corner of the room, as well as what appeared to be the ashen remnants of a small fire. One interviewed witness stated that a fire had been lit in the room after the alleged incident, reportedly to detoxify it of the alleged chemical.

The first video recorded at this site, timed at 10:06 pm, shows that the soot from the fire, no longer burning, has blackened the wall and the ceiling, but that the cylinder appears to be frosted from autorefrigeration. From this and subsequent images it is evident that the soot on the cylinder was beneath the frost: in other words the blackening of the cylinder preceded the autorefrigeration. The cylinder valve had been broken off or removed by the time that images of the cylinder in daylight were recorded. Its absence from the site is unexplained.

If the cylinder was lying on its side or tilted downwards when the valve was first opened, removed or broken off, it would have squirted out liquid chlorine until the level of liquid in the cylinder fell below the valve opening. As the cylinder released gas generated by boiling of chlorine, autorefrigeration would begin: to cool a steel cylinder weighing 64 kg from 24 ℃ (the recorded temperature at Damascus airport at 8 pm) to -34 ℃ (the boiling point of chlorine at atmospheric pressure) would have required evaporation of about 6 kg of liquid chlorine. After this, the rate of discharge of the remaining chlorine would have been limited by the rate of heat transfer to the cylinder from its surroundings. In a reported experiment to estimate the maximal rate of gas withdrawal from a 150 lb chlorine cylinder in still air at 21 ℃, the rate of discharge after autorefrigeration had reached a steady state was about 4 kg/hour, implying a heat transfer rate of 0.32 kW across a 55 K temperature difference.

The witness’s statement that the fire had been lit after the alleged incident is not consistent with the temporal sequence inferred from the images: fire causing blackening of the ceiling and the cylinder, followed by autorefrigeration. It has been suggested that the fire was lit in order to melt the fusible safety plug. Chlorine cylinders have such a plug on the rear of the valve, designed to melt and release pressure in a relatively safe manner if the temperature exceeds 70 ℃.

4.3 Victims did not attempt to escape

The report notes that:

videos taken inside the building indicate exposure to a rapidly incapacitating or a highly toxic substance. The victims do not appear to have been in the midst of attempting self-extrication or respiratory protection when they collapsed, indicative a very rapid or instant onset. This type of rapid collapse is indicative of an agent capable of quickly killing or immobilising.

As reviewed in our earlier briefing note, experts who commented in April 2018 had agreed that these images were compatible with a nerve agent but not with chlorine. In industrial accidents with chlorine those exposed do not drop dead on the spot but usually manage to escape. The build-up of gas in the apartments below would have been limited by autorefrigeration of the cylinder, as noted above, and by escape of gas through the windows of the stairwell; the report notes that “Each level on the staircase has a tall glass-shattered window facing onto the street.”

In reported incidents of chlorine poisoning, non-fatal cases requiring hospital treatment typically far outnumber fatal cases. For instance the Annual Report of the American Association of Poison Control Centers for 2016 [Table 22(A)] shows that in that year 3845 cases of poisoning by chlorine gas were reported in the US, of which 1175 were treated in a health care facility and two were fatal. The case fatality rate of the incident at Location 2, with 43 on-the-spot fatalities documented by images but no reports of cases requiring prolonged hospital treatment, is unlike any recorded incident of chlorine poisoning.

The FFM notes that images at Location 2 showed that many victims had wet hair, which it tentatively attributes to diaphoresis (profuse sweating):

Many of the victims appear to have wet hair in what seems to be an otherwise dry environment.

The presentation of wet hair in an otherwise dry environment is difficult to assess and is possibly due to profound diaphoresis shortly before death.

This does not explain why only the faces and hair of the victims are wet, or why they remain wet after their supposedly sudden deaths which the report assigns to a time three hours before the images were recorded. Diaphoresis is not a feature of chlorine exposure. A possible explanation suggested elsewhere, consistent with the hypothesis that victims had been killed somewhere else, is that victims’ faces were washed after death to remove signs of how they had been killed.