News & Investigations

News & Investigations


April 7, 2022

Written by

Chris Woods

Airwars speaks to Uladzimir Shcherbau, head of the UN civilian casualty monitoring team in Ukraine, on the challenges of tracking civilian deaths.

Beginning in 2014, the United Nations Monitoring Mission in Ukraine has been tasked with recording civilian casualties in that country. Until February 24th, the conflict was relatively low intensity. Russia’s invasion and subsequent attacks on communities across a swathe of Ukraine saw a huge rise in reported civilian deaths, which the UN Mission continues to track.

With the Ukraine Government ceasing its own public national estimates of civilian harm just four days into the war, the UN’s own daily cumulative tallies have become the sole official source for national casualty estimates. By early April, the UN had recorded at least 1,500 civilian deaths – while also publicly acknowledging that was likely a significant undercount.

Uladzimir Shcherbau joined the UN Monitoring Mission in Ukraine back in 2014, and today leads its civilian casualty monitoring unit. Almost all of the team remains in Ukraine – some close to the front lines – and all of them doing extremely challenging work.

In this extended interview with Airwars director Chris Woods conducted on April 1st, Uladzimir discusses among other topics the challenges of UN casualty counting following Russia’s invasion; how the Mission plans to address its own low estimates of harm; and the horrors of Russia’s onslaught on the Ukrainian city of Mariupol.

Some key points from this interview were also summarised in this article for The Independent: UN estimate of civilian casualties in Ukraine set to increase, official says

What’s the purpose of the UN gathering civilian casualty data? 

Why do we do it? It’s not just about gathering data. We have three or four major objectives, to which the information we collect should serve. The first is decision making by military and political actors – ideally to facilitate the cessation of hostilities, and definitely compliance with International Humanitarian Law (IHL). When we see the patterns of IHL violations – it is not our major objective, and very often in such contexts it is very difficult to establish whether the civilian was the result of IHL incompliance or in the worst cases of war crimes – but surely some of our findings would inform the deeper investigations and we do work deeply in some incidents – as we did for the period 2014-2021.

Ideally we go really deep into each single civilian killed, we establish as an ideal minimum the date of the incident, the status of the victim killed or injured, then if possible the name of the victim, then the age, definitely sex, place where it happened – as precisely as possible, if not geographic coordinates – then ideally the settlement or a specific area. Then control over the area – which party controlled the area – and then the weapon and the situation in which the casualty occurred. And we only record civilian casualties, not military casualties. When we have the slightest hesitation that a person might have been Directly Participating in Hostilities, we don’t include them in our figures.

In the current context I would put it like this. Because the amount of material we have to process is enormous and the time constraints are also enormous, we cannot go too deeply into each individual incident. I would describe this as ‘temporary superficiality’.

How is UN civilian harm monitoring presently structured in Ukraine? How big is the team? And are you still able to operate?

Most of the team is located in Ukraine in various places – not in the places where the hostilities are going on but in other places, sometimes nearby, which allows us to reach out to Internally Displaced Persons (IDPs), to people who evacuated from the conflict-torn areas, and to get access to first hand witnesses and victims. And we are still physically present in Donetsk and Luhansk, which Russian-affiliated armed groups control. We have been in those areas since 2014 and 2017 respectively, and we are still physically present in Donetsk, with certain restrictions on our operations there, and also those stemming from the security situation.

We have 37 human rights officers and several interns who have been working on monitoring and to a certain extent everyone is engaged in civilian casualty recording as well. But dedicated work on civilian harm is done by a small unit which I am in charge of.

From 24 Feb—5 April, we recorded 3,776 civilian casualties in context of Russia’s armed attack against #Ukraine: 1,563 killed, incl 130 children; 2,213 injured, incl 188 children, mostly caused by shelling & airstrikes. Actual toll is much higher. Update

— UNHumanRightsUkraine (@UNHumanRightsUA) April 6, 2022

How have you been able to continue this critical work following Russia’s invasion in February?

We apply the global [UN] guidance on how to record civilian casualties and it is fairly simple – we collect information from the best sources we can, preferably at least three independent sources and then based on the body of information we collected, we assess the sources based on their credibility and reliability. We also follow the patterns of violations which also allow us to assess, to verify, and to corroborate – so we collect a lot of information about the hostilities in general which also allow us to better understand the specific incident.

Basically we have three or four types of information. Definitely we take stock of all available official information – local authorities, national authorities, regional authorities, medical institutions, emergency services, police, very often military experts and some others – any state official or a person working in a facility who has some information on casualties. We take stock of all the information they provide.

We often reach out with requests for specific information but in the current context we see that the state agencies are overwhelmed with the numerous challenges, and they don’t have time to respond to our questions. But they do still report on civilian casualties – for example giving totals per region. Not only figures but we also receive a lot of descriptive narrative for specific incidents.

And when someone says ‘ok there was an incident in settlement X and there were a lot of civilian casualties’, in our records, as you can understand, it would be recorded as zero casualties unless we can understand the specific incident. But we take stock of all this information, it is extremely valuable, and we do analyse it systematically. So we have a mosaic of all these official reports.

Within weeks we saw which of these authorities provide consistently reliable information, and those which unfortunately are giving extremely vague information based on plausible assumptions and these assumptions do not get additional credibility simply because they are made by an official. They are definitely free to make any assumptions or judgements based on intuitive thinking but we clearly see in some cases it is not factually based.

The second source for our team is publicly available information on these incidents. In that regard the situation in Ukraine is extremely transparent – available through publicly available sources – Telegram, Facebook, Internet. Roughly speaking, we are talking about a lot of video footage, a lot of photos, and a lot of narrative reports. The number of channels in which people report is enormous. We do systematically follow as much as we can on these channels – very often it’s a local channel for a specific village or something. If you think about Mariupol – which is besieged and the situation is really bad and information is imprecise – people are reporting per quarter what happened in a specific house or area. We use all this information for our analysis and we collect it systematically.

Then there is our outreach to informed individuals. We had, for example, developed a broad network of contacts prior to the present high intensity conflict. All our offices and field presence were mostly in the east of Ukraine. And we are reaching out to these people to get information that is extremely valuable. I will give you just one example – we had a trusted partner NGO which had worked a lot in the east and there were for instance two towns in the Lugansk region which were well known to those who follow Ukrainian events. Shastiya and Staniska-Uganska. And there were hostilities from the very beginning there, and the NGO reached out to medical professionals working in the area to get very detailed medical records about casualties. So we use this network of contacts. They proactively provide us with some information but we also request them to provide us with some information.

And finally, we also publicly announced that we are interested in receiving information about civilian casualties – and so people send us information through Telegram, Facebook and by email.  We cannot cover it (fully) as we don’t have the capacity but we increasingly interview IDPs from the conflict affected areas – and also receive first hand accounts of individual incidents in which civilians were killed or injured.

So altogether we collect this information systematically – not only what happened yesterday but about what happened three weeks ago. If there is a (new) photo of a grave with the names on it from the town recently retaken by Ukrainian forces we go back to the records and check the names. All of that together.

Then we analyse this information, and once we come to conclusions that we have reasonable grounds to believe it happened, then it goes to our figures – the case is corroborated. It doesn’t mean we have all the information – maybe we don’t know the sex of the victims or something – but we still believed it happened and it goes into our figures.

Those cases which have not yet reached this stage we call ‘yellow cases’, which are still pending further verification and corroboration and we have many such cases which are in the process of corroboration. Some of them could stay marked as pending corroboration for months, if not years, some of them could be corroborated tomorrow if new information comes. So we are always reassessing such cases and seeing if anything is new. It’s like a mosaic – sometimes you get new information, and the puzzle  comes together and a case becomes corroborated. This is how we do it.

An elderly woman clears the rubble of her house in the aftermath of a Russian airstrike in the village of Ulica Szkolna, Kyiv Oblast, 29 March 2022. 📸 epa / @AtefSafadi #Ukraine #War #Ukrainewar #epaimages

— european pressphoto agency (@epaphotos) March 29, 2022

The volume of casework you are dealing with must be staggering. What adjustments have you had to make?

It is not a dramatic shift compared to what we did before. We did roughly the same from 2014-2021 – the only thing was the intensity of hostilities was much lower starting from 2015 and we then had the luxury (of) following each individual incident really deeply – going to the place, speaking to the people, getting forensic records etc

And today? Let’s say there is a report of an airstrike in Kiev with reported civilian casualties. Then we take note of this report. Then the city council says it was a centre for the distribution of humanitarian aid and 20 people have been killed on the spot – without specifying whether they were civilians or militants. That’s ok but it is not enough for us. Then video footage becomes available, and in that footage we see four bodies of people who appear to be civilians, some of them women, some can’t be identified but they don’t look like military or paramilitary. And there are also some people in the local chat groups saying a ‘horrible thing happened in this street and I was there and it was a Hell.’

So once again having analysed all this information we believe we have reasonable grounds to believe that the incident happened, and the overall context that the city was under fire, the military actors reported that incident, the General Staff is reporting that there was heavy shelling of this city, we have records from previous days that this area is being targeted, probably we will get some satellite imagery showing the neighbourhood affected. We can reach out to our partner in that area and ask if there was something and they might say “I don’t know what happened, I was 5km from that place, but there was something horrible.” And then, we say we saw five bodies. We don’t write ‘many’, we don’t write 20, we write five people were killed in that incident – including perhaps one man and two women – on that day.

This is roughly how we arrive at a conclusion, but once again it depends on the situation

From our experience usually reports of civilian casualties are not fake – sometimes there could be deliberate attempts to present certain incidents, to portray the other side as a perpetrator. But it is so easy to debunk such fakes so I don’t believe it happens often. Mostly reports are accurate, there could be some sort of natural imprecision or mistakes – very often when you receive a report about a civilian and the name – how the family name is spelled may not be found in the registration of population. But then you realise this family name is accurate –  it was just a (spelling) mistake in the report – and you realise it’s indirect evidence that the report is authentic, because most likely the report was made on the phone and the person who wrote down the family name just didn’t hear it properly.

UNOCHA conflict map of Ukraine for April 4th 2022.

Civilians have been urged by some to take up arms during the present conflict and to Directly Participate in Hostilities. How does that affect categorisations for you, when you’re making a determination of civilian harm?

Methodologically it is fairly simple – civilians are those who don’t take direct participation in hostilities. You don’t have to be a member of the armed forces or paramilitary formations to be counted as not a civilian – it could be purely a civilian who took his or her arms to fight. As soon as you directly participate in hostilities you are not a civilian any more. If such people are killed or injured we exclude them from the UN count.

There could be some marginal cases where there could be questions and you cannot be 100% sure how to classify this person as a civilian or not, but statistically these cases are extremely insignificant and for methodological purity we exclude such debatable questions from civilian casualties.

And we have a fairly clear definition. If there is a policeman who is performing regular police functions in the conflict zone and is killed or injured by hostilities and he or she is not taking part in hostilities – just patrolling the area in the conflict zone when he or she was killed or injured by shelling – we consider them a civilian casualty. If a police officer is present in a conflict zone to maintain law and order as part of enhanced police deployment, and basically also controlling the area recently taken under control, or doing some sort of paramilitary function, he or she is not a civilian any more.

We even excluded cases which could be very marginal – for instance if there is a military unit which is stationed in a certain location and there is a civilian who regularly brings some food to this unit – maybe paid or not, but somehow helps them to sustain. It is easy to argue that this person is not taking part in hostilities but we exclude such cases.

Statistically such cases are marginal but we exclude even very marginal cases so probably decreasing the toll a little. But statistically, it’s pretty clear whether a person participates in hostilities or not.

What is the situation for civilian harm in breakaway and Russian-occupied areas? How able are you to capture that?

For us it doesn’t make any difference – we are covering the entire territory of Ukraine within its internationally recognised borders, and we do follow all the reports and do our best to corroborate all civilian casualties wherever they happen. And we do it in regard to territory controlled by the Russian-affiliated armed groups or the territory which is currently under the control of the Russian armed forces. We have certain difficulties in getting the information but it is roughly the same as getting the information from government-controlled territories. We also do breakdowns for Donetsk and Luhansk regions, for government controlled territory, and for territory controlled by Russian affiliated armed groups.

We haven’t yet published figures from the places which had been controlled by Russian Armed Forces – say from the Kharkiv or Chernihiv or Kherson regions. But we are also coming with those figures fairly soon. So once again, we do take stock of all civilian casualties, including in the territory which is presently outside of Ukrainian government control.

In your own daily press releases you are very clear that the true casualty figure is likely much higher. How wide a gulf might there be between UN estimates and the actual civilian toll?

I wish we had this interview in the next week or two. We are working right now on a realistic estimate of the actual death toll of the conflict. We have a big mass of information which allows us to triangulate or somehow approximate the actual death toll. I wouldn’t give you specific figures right now because it is extremely sensitive and we are under enormous pressure because we are criticised heavily – they are saying ‘your figures are irrelevant. It’s nice that you give these at least figures but they are irrelevant because the real death toll is higher’ as we ourselves point out in our daily updates.

But we have data, we have information, we have a really solid methodology – it is not pure maths, it is looking at patterns, correlations, available information per region and per type of casualties to come (up) with this realistic estimate and once again we will soon come up with an estimate fairly soon.

We believe it would be pretty accurate. The figure would be fairly reflective of the actual scale – it will be also conservative, not to go too broad, but we believe it will be fairly close to the actual death toll. All in all we would come with a rather accurate total and the gaps within the estimate of actual total and real total would not be so big. We believe we will give a gist of what the scale is – of the true scale we believe of the casualties – especially of those killed.

To put things in some perspective, the official civilian death toll after the first 6-8 weeks of the US invasion of Iraq was more than 8,000 — due to "Shock and Awe."

The civilian death toll in Ukraine is just over 1,000. It's all hideous, but calls for *WW3* require *sobriety.

— Glenn Greenwald (@ggreenwald) April 3, 2022

Low UN estimates have been cited to downplay deadly effects of Russia’s invasion.

The city of Mariupol reports more than 5,000 civilian deaths so far. What do you make of that estimate?

When we speak about Mariupol, it is definitely the deadliest place in Ukraine currently. We have been in touch with city authorities, with the city council – we do follow the data they provided and we requested additional data and they committed to providing us as much as they can, but they have been working in extremely challenging conditions as you can understand. Even so, they estimate 5,000 civilian deaths from just three weeks of hostilities in March.

Our own UN estimate should come hopefully fairly soon as we managed to collect information about Mariupol, we analysed all the narrative reports, official figures and official reports from medical (authorities) – sometimes fragmented but also some information which allows us to see the pattern from emergency services and video footage. We also do satellite imagery analysis – not only on damage. So we hope to come with a rather realistic, from our perspective, estimate.

The amount of conflict death is enormous, but also surely people kept dying in besieged Mariupol because of regular mortality, many people died as indirect casualties because of stress and the collapse of medical aid. And reportedly many suicides have occured in Mariupol. We have had several reports saying that the suicide rates increased, that’s for sure.

So it needs also to be factored in that all these people died during this one month of hostilities – though the UN Mission ourselves strive to single out the civilians who have been directly killed by hostilities, and we will come up with our estimate pretty soon.

Given this scale of harm, how long might it be before the UN has a comprehensive tally of civilian deaths at Mariupol?

When will the world know for sure when hostilities are over the exact number of civilians who have been killed in Mariupol from day one until the last day of hostilities? That will surely take time. It will take time to recover all the bodies, to identify them, and because there was a mass displacement for example, some people who were evacuated from Mariupol, some injured people who could have died in medical facilities outside of Mariupol. So the ultimate accurate figure won’t arrive quickly.

We have credible reports that there are still many bodies in the debris. Some bodies are unattended in the apartments – no one was (able) to take care of the bodies. Many bodies have been buried in improvised graves – they also need to be exhumed and reburied individually and with proper decency. That will also be an enormous challenge for the city.  We have seen a lot of footage of graves in peoples’ yards, which is appalling.

So we will analyse all the information and try to come up with our own estimate but surely for the future that will be a horrible and heartbreaking task – not only to count people, but to ensure decent treatment of those who perished. And then once again to work for reparations and bring the perpetrators to justice. The preservation of evidence – which is what dead bodies are – will be extremely essential.

Маріуполь: відео з дрону та знищений драмтеатр зсередини

— Радіо Свобода (@radiosvoboda) April 5, 2022

Radio Svoboda’s recent drone footage showing the devastation of Mariupol.

▲ A woman passes bodies buried at the side of the road during Russia's siege of Mariupol, April 3rd 2022 (Reuters)


December 18, 2021

Written by

Chris Woods and Joe Dyke

Almost 800 previously secret US airstrikes in Afghanistan during 2020 and 2021 are revealed, as US military declassifies data.

The release of classified records of recent US airstrikes in Afghanistan has revealed more than 400 previously undeclared actions in the last months of Donald Trump’s presidency – and at least 300 more strikes ordered by Joe Biden’s administration.

Even after the United States and the Taliban signed an effective peace agreement in February 2020, the US continued secretly to bomb Taliban and Islamic State targets, the data shows. And during 2021 – as the Taliban continued to ramp up attacks on Afghan government forces, and advance on Kabul – more than 800 munitions were fired by mostly US aircraft.

The crucial Afghanistan monthly data by Air Force Central Command, or AFCENT, was stopped in March 2020 after the Trump administration agreed an effective ceasefire deal with the Taliban. Those public releases showed how many strikes the US and its international allies carried out in Afghanistan as well as details of weapons fired, and had been released monthly for nearly a decade beforehand.

At the time the US Air Force said it was stopping the releases due to diplomatic concerns, “including how the report could adversely impact ongoing discussions with the Taliban regarding Afghanistan peace talks”.

The newly declassified data adds credence to allegations at the time that the United States may have secretly upped its strikes in Afghanistan to put pressure on the Taliban during negotiations taking place in Qatar, with sometimes devastating impacts for civilians.

While the United Nations was seemingly convinced that US strikes had largely stopped, the Taliban accused the US of violating the terms of the agreement “almost every day.” Those claims are now more likely to be taken seriously.

“These data tell the story of America’s struggle to end its longest war,” Graeme Smith of International Crisis Group told Airwars.

An air war that never ended

The US and the Taliban signed a so-called ‘peace’ arrangement on February 29th 2020. This did not explicitly commit the US to a full ceasefire, but involved the Taliban effectively committing not to attack American forces in Afghanistan during a proposed 14-month US withdrawal period.

It was also assumed that US strikes would also significantly wind down, and be focused primarily on self-defence actions. Yet the newly released AFCENT data shows US attacks never ceased, with 413 ‘international’ airstrikes between March and December 2020 alone.

Declassified AFCENT data has revealed almost 800 previously undeclared airstrikes conducted in Afghanistan during 2020 and 2021

Following the US-Taliban agreement in February 2020, official ceasefire talks then began in Doha in September of the same year between the Taliban and the Afghan Government. Yet in the same month, we now know, the US still secretly conducted 34 airstrikes.

Continuing US actions coincided with Taliban onslaughts on the outskirts of the cities of Kandahar and Lashkar Gah. The Taliban argued that these assaults, on Afghan government forces rather than American ones, were not in breach of the agreement but the US disagreed, Smith said. “That is why you see a sharp uptick in airstrikes from October 2020 as the Americans desperately tried to defend those provincial capitals,” he said.

Amnesty International recently highlighted what it believed was a US airstrike on Kunduz in November 2020 which killed two civilian women, Bilqiseh bint Abdul Qadir (21) and Nouriyeh bint Abdul Khaliq (25), and one man, Qader Khan (24). Munition fragments recovered from the scene pointed clearly to a US strike. It is now clear that the United States secretly conducted 69 strikes in Afghanistan that month alone.

Since assuming office in late January 2021, Joe Biden initially oversaw a slight drop in strikes before a significant increase, as the 20-year US occupation ended in a chaotic and devastating withdrawal.

In the final desperate three months of the US presence, 226 weapons were fired in 97 airstrikes by US (and possibly allied) aircraft in a doomed bid to halt the Taliban’s lightning advance. Many of those actions were likely to have been close air support strikes aiding Afghan National Army forces in urban areas, who were being overrun. The known risk of high civilian casualties from such actions has long been known.

In the chaotic last days of the war, dozens of civilians and 13 US military personnel died in an ISIS-K suicide attack as US forces barricaded themselves inside Kabul airport and desperate Afghans flocked to the site hoping to flee the country.

And in the final airstrike of the US occupation, 10 civilians were killed when American drone operators confused a father returning to his family home with an Islamic State terrorist. Last week, the Pentagon announced no disciplinary action would be taken in that strike.

United Nations deceived?

Stopping the release of monthly airstrike data in early 2020 also appears to have convinced the United Nations that the US was no longer conducting significant attacks.

In both its 2020 annual report on civilian casualties in Afghanistan and its 6-monthly report for the first half of 2021, the United Nations Assistance Mission to Afghanistan (UNAMA) played down the impact of US and international strikes – believing them to have mostly ended.

During 2020 the UN concluded, more than 3,000 Afghan civilians were killed in ongoing fighting between the Taliban and the then-Afghan government, supported by international forces. According to UNAMA, 341 civilians were killed that year by airstrikes – of which it blamed 89 deaths on international forces.

Yet UNAMA’s 2020 annual report said that after the February 29th agreement between the US and Taliban “the international military significantly reduced its aerial operations, with almost no such incidents causing civilian casualties for the remainder of 2020.”

UN officials later told Airwars during a briefing that they believed Afghan Air Force strikes were now likely responsible for almost all civilian deaths from airstrikes. The release of the previously-classified data from AFCENT radically changes that picture. Between March and December 2020, Trump’s last full months in office, the US in fact carried out 413 airstrikes – as many as during all of 2015 for example.

For the first half of 2021, UNAMA also made similar assumptions about low numbers of US and international strikes, noting that “compared with the first half of 2020, the total number of civilians killed and injured in airstrikes increased by 33 per cent. Civilian casualties from Afghan Air Force airstrikes more than doubled as international military forces conducted far fewer airstrikes.”

In fact, we now know, more than 370 ‘international’ strikes were carried out in 2021, which between them dropped more than 800 munitions.

UNAMA did not immediately respond to questions on whether the UN would now be reviewing its recent findings, following the release of the AFCENT data.

Biden under scrutiny

Revelations of hundreds of previously secret US airstrikes in Afghanistan during Joe Biden’s first months in office indicate that while US actions were at record lows in other theatres such as Iraq and Somalia, the intensity of the 20-year war in Afghanistan continued to the very end.

More than five times more US strikes were conducted in Afghanistan from January to August 2021 than have been declared in all other US theatres combined across the whole year, Airwars analysis shows.

“Airwars has been cautioning for some time that recent airstrike numbers for Afghanistan – if revealed – might show far more US military activity under Joe Biden than many had assumed,” said Airwars director Chris Woods. “This newly released data – which should never have been classified in the first place – points to the urgent need for reevaluation of recent US actions in Afghanistan, including likely civilian casualties.”

The Afghanistan data stops abruptly in August 2021. Announcing the release of the previously secret strike and munition numbers to the Pentagon press corps late on Friday afternoon, chief DoD spokesman John Kirby told reporters: “There have been no airstrikes in Afghanistan since the withdrawal is complete.”

▲ A home allegedly destroyed by a US airstrike in Kunduz, Afghanistan in November 2020 (Credit: Amnesty International)


December 2021

Written by

Chris Woods and Emily Tripp

Assisted by

Adam Gnych, Ayana Enomoto-Hurst, Clive Vella, Dmytro Chupryna, Duncan Salkovskis, Edward Millett, Georgia Edwards, Hannah Aries, Imogen Piper, Megan Karlshoej-Pedersen, Mohammad Al Halabi, Sanjana Varghese, Shihab Halep and Valentina Finckenstein

‘Why did they bomb us?’ Urban civilian harm in Gaza, Syria and Israel from explosive weapons use

‘Why did they bomb us?’ Urban civilian harm in Gaza, Syria and Israel from explosive weapons use’ – which is published in Arabic, Hebrew and English – provides the first comparative analysis of two very different Israeli military campaigns.

Applying the same open source monitoring methodology Airwars uses to track civilian harm caused by the United States, Russia, and other actors in conflict nations, researchers documented all local reports of civilians killed and injured in May 2021 by Israeli strikes in Gaza as well as of civilians harmed in Israel by Palestinian rocket fire. The report also examines locally reported civilian casualties from Israel’s eight-year long Syria campaign against Iranian-linked forces.

High civilian casualties in Gaza are symptomatic of an escalating and profoundly troubling global military trend in the use of wide area effect weapons in populated areas. Airwars’ report demonstrates that choices made by belligerents continue to have devastating effects on civilians – clearly showing why the use of explosive weapons in urban centres must be restricted.


November 18, 2020

Written by

Chris Woods

Assisted by

Abbie Cheeseman, Alex Hopkins, Clive Vella, Eeva Sarlin, Hanna Rullmann, Ned Ray and Sophie Dyer

Major transparency breakthrough may help Iraqis and Syrians to secure restitution and reconciliation

The US-led Coalition has released to Airwars the near coordinates of almost all confirmed or ‘Credible’ civilian harm events in Iraq and Syria in the long war against so-called Islamic State, allowing for the first time the accurate locating of 341 confirmed incidents and almost 1,400 civilian deaths since 2014.

The groundbreaking decision by the US-led Coalition – which came after several years of patient engagement by Airwars – will enable affected Iraqis and Syrians for the first time to know whether their loved ones were caught up in a particular event. That in turn could open the way for both apologies and ex gratia payments from the US and its international allies.

Former chief Coalition spokesman Colonel Myles Caggins said that the decision to share the data had been taken in the interest of transparency: “We take every allegation of civilian casualties with the utmost sincerity, concern, and diligence; we see the addition of the geolocations as a testament to transparency, and our commitment to working with agencies like Airwars to correctly identify civilian harm incidents.”

“The release of this locational data for confirmed civilian harm events in Iraq and Syria – accurate in some cases to just one metre – sets a new and welcome transparency benchmark,” Airwars noted. “We appreciate the US-led Coalition’s decision to release this important material, which should help affected Iraqis and Syrians to secure some closure following tragic losses within their families.”

What the data shows

In August 2014 the US-led coalition began bombing so-called Islamic State after militants had seized large swathes of territory in Syria and Iraq. Millions of civilians were trapped in a brutal war that lasted several years, with fighting often taking place in heavily built-up neighbourhoods.

Over the years the US-led Coalition has admitted many hundreds of civilian deaths from its own actions – though identifying exactly where these took place has been a major challenge.

The release to Airwars of hundreds of coordinates of Credible incidents – most accurate to within 100 metres and some to within just one metre  – is believed to be the most comprehensive locational civilian casualty data ever released by the US military.

Airwars has now added a new mapping and research tool to its website, The Credibles, which comprehensively maps all located incidents across both Iraq and Syria. The data has also been visualised in partnership with The Washington Post.

The locational data provided by the US military is, Airwars believes, unique. No previous belligerent is thought to have revealed at scale – either during or after a war – exactly where and when it has harmed civilians.

Using the US Department of Defense’s preferred Military Grid Reference System (MGRS), 70 of these Credible civilian harm events have now been publicly geolocated by the Coalition to an accuracy of just one metre squared, with all additional events geolocated to an accuracy of a one hundred metre square box.

Many of these incidents have already been well documented by affected communities themselves, with associated photographs, videos, and eyewitness narratives. Airwars also presently lists the names of more than 900 victims from these located events.

Just three of 344 confirmed Coalition civilian casualty incidents have been omitted from the data release. One recent case involved Coalition ground troops in Iraq, suggesting possibly sensitive Special Forces operations. Another event is still being queried with US Central Command (CENTCOM). The third case is the sole British-confirmed anti ISIS event, from March 2018 – with the UK Ministry of Defence still declining to release any locational information.

Some of the 341 Credible locations released to date by the US-led Coalition

The civilian casualty assessment process

More than 29,600 civilians have locally been alleged killed by US-led Coalition actions in Iraq and Syria since 2014, according to Airwars monitoring of local populations. The US was initially slow to engage – with just three civilian harm events confirmed by CENTCOM in the first 16 months of the war.

Beginning in 2016, the process of casualty assessments by the US military became more systematised – in part as a result of an increased focus on casualty mitigation by the Obama administration during its last months in office; and in part because of pressure from Airwars and other NGOs, which between them were tracking local allegations of civilian harm from Inherent Resolve actions in both Iraq and Syria.

CENTCOM established a permanent civilian casualty assessment team at Tampa covering the war against ISIS in mid 2016, and began publishing more regular public reports on confirmed civilian harm events. An additional 60 Credible incidents were admitted that year, for example.

In December 2016, the Coalition took over civilian casualty assessments from CENTCOM (although almost all personnel continued to be drawn from US forces.) The US-led alliance also began publishing monthly civilian harm assessments – which continue today. In total, CENTCOM and the Coalition have now assessed almost 3,000 alleged civilian harm events in the war against ISIS, to date confirming as Credible some 344 of these incidents.

Around one in four of these Credible events have resulted directly from Airwars referrals to the Coalition, meaning that without local reporting by affected communities – and the patient work of Airwars’ own Syrian and Iraqi researchers collating and preserving those reports – the events would not have come to light.

The challenges of properly locating Credible events

While the confirming of multiple civilian casualties in Iraq and Syria by CENTCOM and the Coalition was generally well received, there were often significant challenges in properly identifying the location of such events.

Press releases issued via military public affairs teams only tagged a Credible event to the nearest large population centre. In the September 2017 report for example, the Coalition declared a major confirmed civilian harm event in Mosul: “March 14, 2017, near Mosul, Iraq, via media report: During a strike on ISIS fighters engaging partner forces from a fighting position, it was assessed that 27 civilians in an adjacent structure were unintentionally killed.”

The month of March 2017 saw very heavy fighting at Mosul. Airwars tracked five separate claimed civilian harm events in the city for March 14th alone (two of them mass casualty incidents), with some days seeing more than a dozen allegations. Without the correct coordinates, affected Moslawis could never know whether their loved ones were (or were not) affected by particular strikes.

For the March 14th 2017 event cited above, Coalition officials eventually provided Airwars with coordinates in Mosul accurate to within just one metre (38SLF2901422174), confirming that the event took place in the neighbourhood of al Jadida. The challenge was whether such precise data could now be acquired for all confirmed events.

Such data could also help to prise open the door on possible restitution for civilian harm from US and Coalition actions in Iraq and Syria. According to the Pentagon, only six ex gratia payments were made to Iraqis during 2019, compared with more than 600 such awards for Afghanistan. Poor public locational data from Operation Inherent Resolve for confirmed civilian harm events has majorly hampered the ability of Iraqis and Syrians to pursue claims – until now.

Ex gratia payments awarded by the US in Afghanistan during 2019 were one hundred times higher than in Iraq. Poor public locational data by CJTF-OIR for confirmed events likely contributed to that disparity.

Securing the locational data

Following a face to face meeting with senior officials in Tampa in May 2016, CENTCOM and Airwars began regularly sharing data on civilian casualty allegations, in order to improve understanding, on both sides, of reported non combatant harm. That relationship has continued, with sometimes weekly confidential engagements between the Coalition’s CIVCAS Cell and the Airwars military advocacy team, involving granular queries from both parties.

CENTCOM also began sharing with Airwars occasional precise locational data for Credible events in mid 2016, in order better to clarify particular cases. Over time this became more systematised.

Alongside its monthly public press releases, the Coalition for several years provided Airwars with a private, annotated version of the monthly release, which both identified the geolocation of the event – and also, where possible, specifically cross matched Credible incidents to coded events already in the Airwars database. This locational information was provided by CJTFOIR on the expectation that Airwars would make it public through its own database of civilian harm events.

By early 2018 the Coalition was consistently providing MGRS data every month to Airwars for both Credible and later, for ‘Non Credible’ events. However this still left 126 historical confirmed cases for which Airwars had no locational data.

In early 2019, Airwars asked both the Coalition and the US Department of Defense to release this information – arguing that the alliance’s significant transparency in confirming civilian harm cases was being weakened by our then being unable publicly to determine where such cases had actually occurred. It was also argued that the US could better distinguish itself from Russia and other belligerents, who instead chose to hide or deny civilian harm from their own actions.

The missing locational data was provided to Airwars in Summer 2019 by the Coalition’s civilian casualty assessment team – a major step forward for transparency and public accountability. Later that year, the Coalition also began publishing MGRS data as part of its regular monthly public reporting.

The Coalition began sharing Credible close coordinates with Airwars in 2016, at first in private annotated versions of public reports.

Visualising the data

The Credibles dataset offers significant potential for visualisations, allowing for each confirmed US and allied civilian harm event in both Iraq and Syria to be precisely mapped and timelined. Additionally, each event can be cross matched to an associated Airwars incident report. More than 900 victim names are linked, along with associated photographs, videos, witness statements, and satellite imagery analysis of bomb sites.

Airwars has built a new subsite illustrating the remarkable potential of this unique transparency dataset. Each US-led Coalition Credible event has been precisely mapped down to at least 100 metres, and timelined across the war.

“The significance of this information for the affected communities led us to create an interface that would make the dataset easily accessible, and represent the information in a way that reflected its accuracy,” says Lizzie Malcolm of Scottish-American design team Rectangle, which conceived the new subsite.

The Credibles data sets a new benchmark for US military accountability for civilian harm. It’s hoped that the release of such accurate geolocational data can now become standard both for the US military and its allies.

▲ The US-led Coalition has released to Airwars the close locations of almost every confirmed civilian harm event since 2014.


October 2020

Written by

Chris Woods, Laurie Treffers and Roos Boer (PAX)

Seeing through the rubble: The civilian impact of the use of explosive weapons in the fight against ISIS

A joint report ‘Seeing through the rubble: The civilian impact of the use of explosive weapons in the fight against ISIS‘ by Airwars and PAX examines the dire and long-lasting effects of explosive weapons on civilian populations in towns and cities, in recent international military campaigns in Mosul, Raqqa and Hawijah.

Explosive weapons kill and injure people upon use, and often have an impact that extends far beyond the time and place of the attack. They are a major driver of forced displacement – not only because of fear of death and injury and the destruction of homes, but also because of their profound impact upon critical infrastructure services such as health care, education, and water and sanitation services.

In order to better protect civilians from the use of explosive weapons in populated areas, the authors of the report call upon States to integrate the direct, indirect and reverberating effects of the use of explosive weapons into their military planning and operations, and to develop and support a strong international political declaration to better protect civilians against the use of explosive weapons in populated areas.


March 15, 2020

Written by

Chris Woods

Assisted by

Abbie Cheeseman

Key European allies are denying dozens of civilian deaths from their own actions - even where the US-led Coalition finds such cases to be credible.

A major international investigation has found compelling evidence that several of the US’s key European allies in the war against so-called Islamic State routinely deny civilian harm from their own actions – even where specialist US military personnel within the international Coalition have assessed such cases to be credible.

Three European countries are implicated – the United Kingdom, France and Belgium – a lengthy investigation by the BBC, Libération, De Morgen and RTL Netherlands has found.

BBC News: US military says strikes may have killed civilians

Libération: Syrie-Irak : ces frappes meurtrières que les Etats refusent de reconnaître

De Morgen: De veertig burgerslachtoffers die niemand erkent, ook België niet

A total of eleven specific civilian harm events have so far been identified – involving the officially confirmed deaths of at least 40 Iraqi and Syrian civilians during 2017 and 2018. No European ally will admit to the fatalities.

“Cases like this expose a fundamental gap in accountability created by multinational coalitions,” notes Dan Mahanty of the US advocacy organisation CIVIC. “If warring parties simply collude to hide their actions, they can also evade their responsibilities. For civilians who lost loved ones or had their livelihoods destroyed, it means losing any hope of remedy, or even basic acknowledgement of their loss. It’s a pretty significant affront to their dignity.”

US admissions

The problem incidents came to light after the US Defense Department was legally required to report to Congress in May 2019, on all recent confirmed civilian deaths from US military actions. That Pentagon report declared 170 incidents for Iraq and Syria during 2017; and a further 13 events during 2018.

However, when Airwars then crossmatched the 183 declared US civilian harm events against those cases the anti-ISIS Coalition had officially conceded during the same period, it identified 14 further incidents which had been omitted. Several senior US defense officials independently confirmed to Airwars that all credible non-US civilian harm events had been explicitly excluded from the list given by DoD to Congress.

Three of these ‘missing’ events were previously confirmed Australian civilian harm cases. That left eleven civilian harm incidents which had not publicly been admitted by any US ally – for example the deaths of three civilians on May 28th 2017 including Hayat, the wife of Mustafa al-Saguri, who died alongside her young daughter and a third unknown civilian at al Hammam in Raqqa province.

The US-led Coalition had admitted those deaths in April 2019, noting that “Regrettably, the strike on an associated target building unintentionally resulted in the deaths of three civilians.” But which US ally was responsible?

During 2017 and 2018, five partners were still active alongside the US in the war against ISIS: the UK, France, the Netherlands, Belgium and Australia. With the al Hammam incident not included in the Pentagon’s report of US-caused civilian harm events to Congress, Airwars then crossmatched this and ten other unclaimed incidents, against published strike reports by the US’s allies for the dates in question.

In June 2019, Airwars wrote to each nation’s military, requesting confirmation of whether its forces had been responsible or not for specific confirmed civilian harm events.

Australia quickly responded that it had not conducted either of the incidents it had potentially been flagged in, noting definitively that “Australian aircraft did not conduct either of the strikes on 9 January 2017 and 15 May 2017.” Nine months later, the Dutch ministry of defence finally confirmed that it was not responsible for those deadly strikes it had in theory been linked to.

With indirect confirmation that the eleven officially confirmed civilian harm events had been the responsibility of three European militaries, Airwars then approached major news organisations with which it had engaged previously on civilian harm issues. The BBC, Libération, RTL Netherlands and De Morgen each then pursued its own national investigation, with an agreed joint embargo.


Britain: admits strikes but denies civilian deaths

The most comprehensive admission during the investigation came from the UK’s Ministry of Defence (MoD), which is among the more transparent members of the US-led Coalition.

Of the six potential UK events flagged to it by Airwars, the MoD confirmed by letter that it had been responsible for three of the strikes which the Coalition assessed had killed at least 15 civilians. However, the UK then refuted the US-led Coalition’s findings, insisting that no civilians had in fact died.

The first event took place at Mosul, Iraq on January 9th 2017, and had already sparked a BBC investigation after a military whistleblower within the Coalition had reported civilian deaths from an RAF strike. The Ministry of Defence had over-ruled that view, determining that no civilians had been harmed.

Following the BBC’s investigation, US military personnel at the Coalition had themselves then assessed the event – determining it to be Credible in March 2018 and noting that “two civilians were unintentionally killed.” In a detailed letter to Airwars, the MoD justified its own continuing refusal to accept civilian deaths:

We looked at this incident very closely indeed. It happened in an area of active fighting between Daesh and the Iraqi security forces, and neither the troops on the ground nor the coalition aircraft detected any signs of a civilian presence in the area either before or after the truck-bomb was destroyed. The group of men, by their movements and behaviour, showed every sign of being Daesh fighters – particularly the presence of a motorcyclist, frequently used by the terrorists to scout ahead during the street fighting…. We therefore concluded that, if the group did indeed sustain casualties, they were extremely likely to have been Daesh terrorists; we have no reason to believe, on the evidence available, that they were civilians.

The second RAF strike took place at Raqqa, Syria on August 13th 2017. According to the US-led Coalition, 12 civilians died after “Coalition aircraft engaged ISIS fighters utilizing a mortar system in a building used as a defensive fighting position.” Among the victims locally named that day were Walid Awad Al Qus and his young daughter Limar.

The Coalition’s admission of 12 deaths in this event represented one of the highest confirmed tallies for the entire battle of Raqqa, which an Amnesty International/ Airwars investigation later concluded had seen at least 1,600 civilians killed by Coalition actions.

Once again accepting the strike but denying the civilian deaths, the MoD asserted: “A single individual was seen on weapons system video moving in the area just prior to the impact of one of our weapons. There is no evidence that this individual was a civilian, as opposed to one of the Daesh fighters engaged with the [SDF]. We have certainly not seen any evidence that twelve civilian casualties were caused.”

In the final event, an RAF drone strike on January 20th 2018 killed one civilian nearby, according to an internal assessment by the US-led Coalition. Once again, the British reached a different conclusion. “Careful analysis was conducted of the available footage and of all available reports from the area. These showed that there was no evidence of civilians being present in the location, and the footage identified a weapon being carried by the likely casualty. It was therefore concluded that said individual was very likely a Daesh extremist and not a civilian.”

Senior defence officials confirmed to both the BBC and to Airwars that the UK presently requires what it calls ‘hard facts’ when assessing civilian harm claims – an apparently higher standard even than the ‘beyond reasonable doubt’ used by UK courts. Civilian casualty assessors within the US military instead use a ‘balance of probabilities’ approach, Airwars understands – allowing them to consider local credible reports of civilian harm in their own investigations.

In effect, the UK has set the burden of proof so high that it is almost impossible for the MoD to reach a determination of civilian casualties – even when its most powerful ally the US concludes the exact opposite, critics say.

Chris Cole of the advocacy group Drone Wars UK accused the Ministry of Defence of overly focusing on managing public perceptions of war, rather than looking at appropriate levels of transparency: “We end up with obfuscation, secrecy and – as these revelations show – a kind of internal structural self-denial, where it has become seemingly impossible for the MoD even to accept that civilian casualties have occurred.”

Library image: Missiles being loaded onto an RAF Tornado prior to a mission against Daesh (via Ministry of Defence)

Belgium: “Certainly not involved in all events”

Belgium, which ended its involvement in the war against ISIS in late 2017, was potentially implicated in at least nine incidents which the US-led Coalition had deemed credible, an initial review concluded. While two of the strikes were later admitted by the UK, at least 23 civilians had died in the other seven events in which Belgium was implicated, according to US officials.

Several of these incidents were already known. In May 2017, a senior Belgian official had briefed Airwars that the Government was planning to admit two civilian harm events in Iraq earlier that year – one at al Qaim on February 27th, and a second event on March 21st near Mosul. Between them, the US-led Coalition had itself concluded, the strikes had killed at least two civilians and injured four others.

However Belgium then failed not only to declare its role in the strikes, but also publicly denied any civilian harm – which led in turn to a front page news story in De Morgen at the time.

Asked in June 2019 to say whether its aircraft were responsible for officially declared civilian harm in up to nine incidents, the Belgian Ministry of Defence told Airwars by email: “For the year 2017, BAF [Belgian Armed Forces] was certainly not involved in all events. With regard to the other data given, BAF was no longer present in theatre. BAF completed its role at the end of 2017. Our conclusion is that all ROEs [rules of engagement] were respected as confirmed by our federal court.”

That comment by Belgium – that it “was certainly not involved in all events”, appears to be tacit confirmation that their aircraft were involved in confirmed civilian harm events. However it remains unclear whether the Ministry accepts the Coalition’s own findings in any Credible case.

In its own investigation, De Morgen features Muhammad Sheikh Sa’ab, whose leg was amputated following a likely Belgian or French airstrike on May 12th-13th 2017. He was one of the lucky ones. The US-led Coalition acknowledges at least 10 deaths, while locals insist more people died. Survivors and relatives may never know which military was responsible.

“Belgium and other Coalition countries cannot bomb and then simply decide to look away from the deadly consequences. If there is proof of civilian casualties, the Belgian government needs to take responsibility,” argues Willem Staes of the Belgian advocacy organisation 11.11.11. “Mature democracies need to ensure both transparency and accountability, and provide civilian victims with adequate compensation and restitution.”

De Morgen interviewed the survivor of a probable Belgian or French airstrike which in 2017 killed at least 10 civilians, according to the US-led Coalition.

The Netherlands: Last minute transparency

When the Dutch government admitted in November 2019 that its aircraft had been responsible for the deaths of approximately 70 civilians in Hawijah, Iraq almost five years earlier – a fact which had been hidden from both Parliament and the Dutch people – defence minister Ank Bijleveld promised new transparency standards. Yet for much of this investigation, it seemed little had changed.

In June 2019, Dutch defence officials were informed of two Coalition-confirmed civilian harm events in which their aircraft were potentially involved.

While one of those events was later confirmed to be a British strike, a second at al Bahrah, Syria on February 9th 2018 still implicated both the Dutch and French militaries. “One civilian was unintentionally killed as a motorcycle entered the impact area moments before the strike,” US military investigators had concluded in August of that year.

In late January 2020, Dutch officials verbally informed Airwars that they would be neither confirming nor denying their involvement in the two Coalition-confirmed civilian harm events.

However, in a last minute turnaround, on March 13th defence officials informed Airwars and RTL Netherlands that they had not, in fact, been involved in either of the incidents, stating that “In the interests of increased transparency, we can now explicitly answer your question about these air raids. As far as the Ministry of Defense is aware, these attacks did not involve Dutch forces.” It was also indicated that from now on, the Netherlands planned to be more transparent in such cases.

“Better late than never, this is a major step in the right direction for Dutch military transparency and accountability. If implemented fully, this should benefit past and future civilian victims seeking information, assistance or compensation and it should benefit parliamentary oversight of Dutch participation in military operations,” says Wilbert van der Zeijden, a team coordinator focused on Protection of Civilians at PAX.

Library image: A Dutch F-16 pilot checking missiles before take-off from an airbase in Jordan (Netherlands Defence Ministry)

France: a refusal to engage

According to an initial review, France was potentially implicated in up to nine Coalition-confirmed civilian harm events during 2017 and 2018. While the UK has since explicitly confirmed its role in several of those strikes, France remains implicated in seven events which between them killed at least 24 civilians.

Despite conducting more air and artillery actions than any Coalition member other than the United States, the French have yet to admit to a single civilian death in their six year war against so-called Islamic State.

That silence continues. After confirming receipt from Airwars in June 2019 of details of possible French civilian casualty events, the defence ministry then ceased communication – refusing to answer all emails ever since.

Marie Forestier, who is part of the Libération team investigating civilian harm from French strikes, previously reported for the newspaper that “200 allegations of civilian casualties potentially involving the French military have been investigated.” Yet details of those investigations remain secret.

Officials do not deny that civilians have been killed by French actions. Even so, they insist that those numbers must remain buried within broader Coalition numbers.

With the United States, the UK and the Netherlands each explicitly denying involvement in a Coalition-confirmed event near al Bahrah village in Syria on February 9th 2018, only France now appears liable. Yet Ministry officials are still refusing to confirm or deny their involvement in the confirmed death of a civilian that day, according to Libération.

“The French Ministry of Defense has refused to answer direct questions and followup questions. As there is a total lack of interest from MPs, media, and public opinion in France, the Army remains unchallenged and is not encouraged to reveal more information. As a result, there is no scrutiny on French airstrikes and no accountability” asserts reporter Marie Forestier, who has been examining French accountability for civilian harm for several years.

Library: French artillery crews in action against ISIS in May 2019 (Image via Armee francaise)

Widening gulf between US and Europe

Airwars is calling for a major review by European powers of their approach to civilian harm assessments – where the US now leads on best practice.

“US military officials are certainly no pushover when it comes to determining civilian harm. Around nine out of ten claimed civilian casualty events assessed by the Coalition since 2014 have been rejected, our analysis shows,” says Dmytro Chupryna, deputy director of Airwars.

“Even so, this investigation reveals a complete unwillingness by most European allies to admit civilian harm from their own strikes – even where US military personnel determine otherwise. Europe’s civilian casualty assessment processes are presently unfit for purpose.”

On May 12th 2017, during the fierce battle for Raqqa, at least 10 and as many as 20 civilians died when Coalition aircraft attacked Asadiya farm, to the north of the city. According to local reports the dead included Khalil Dhammaage; Hassan Ismail Al Zeyabage; Muhammad Al-Nasehage; and Abu Baraa and his entire family.

Three months later, Coalition military officials concluded that “During a strike on ISIS fighters, it was assessed that 10 civilians were unintentionally killed in a building adjacent to the target.” Yet to this day, neither Belgium nor France will say whether their aircraft killed those ten or more civilians. Remaining families have no chance of an explanation, an apology, or compensation.

According to Dan Mahanty of CIVIC, “the record now clearly shows that a public accounting of civilian harm carries few risks and more than a few benefits for belligerents. It’s a shame that the overall record of transparency and accountability for the US-led Coalition is rendered less meaningful because a few governments prefer to hide in the crowd.”

In six weeks, the Pentagon is due by law to make its latest disclosure to Congress on civilian harm claims from US actions, covering a period in which at least 44 additional civilian harm events have been confirmed by the Coalition in Iraq and Syria. How many of these will again emerge as non-US events remains to be seen.

▲ A March 2017 airstrike during the battle for Mosul against Islamic State. While the US-led Coalition has admitted almost 1,400 civilian deaths during the war, European allies have remained almost silent about their own responsibility. (Via Reuters/ Alaa Al-Marjani)


December 19, 2019

Written by

Chris Woods

Airwars analysis of official data indicates US officials were privately acknowledging 70 civilian deaths at Hawijah, long before the Dutch government admitted the event.

The Netherlands Ministry of Defence (Defensie) has provided an inaccurate statement to both MPs and the media, Airwars analysis indicates – on the eve of a critical parliamentary debate on the deaths of scores of civilians in Iraq which had resulted from a Dutch airstrike in 2015.

Credible reports at the time had indicated that at least 70 civilians died at Hawijah in June 2015, after a Coalition airstrike detonated ISIS explosives held in a VBIED factory. Much of the surrounding neighbourhood was destroyed, and the Coalition almost immediately ordered an inquiry into the attack, which has never been published.

However it was only on November 4th 2019 that the Dutch government finally admitted responsibility for Hawijah – following a major investigation by news organisations NOS and NRC.

Since then, as a political crisis has engulfed the coalition government, Defensie has sought to play down or dismiss reports that 70 civilians died at Hawijah – and also that US defence officials had privately conceded those deaths long before the Dutch admission.

Dismissing the evidence

One problem for Defensie was that several pieces of evidence appeared to contradict their denials of 70 civilian deaths. In April 2017 CENTCOM officials had declared 80 non-US civilian deaths, in frustration at their unnamed Coalition allies not accepting public responsibility for events. The suspicion remains that Hawijah formed a significant proportion of those 80 deaths.

Then in December 2018, an on the record email to Dutch reporters from then-Coalition official spokesman Colonel Sean Ryan explicitly stated that “The strike to the VBIED factory caused secondary explosions that unfortunately killed 70 civilians despite the precautions the Coalition took to mitigate civilian casualties.”

Finally, in a declassified 2018 Pentagon report produced by the US National Defense University (NDU) – which was obtained by the Washington Post and published in February 2019 – a graphic appeared to show a clear casualty spike in official US military tallies of civilian deaths in Iraq, at exactly the point at which Hawijah occurred.

With the US-led Coalition recently and inexplicably withdrawing its estimate of 70 civilians killed at Hawijah, Defensie is also now seeking to downplay the importance of the NDU graphic.

In a statement issued to parliament on December 18th, the defense ministry claimed that the NDU graphic did not in fact feature the Hawijah data, asserting instead that “It can be concluded from the table that, based on investigations into possible civilian casualties as a result of the Coalition’s deployment of weapons, CENTCOM was able to confirm [only] that a higher number of civilian casualties occurred in this period than in the preceding and subsequent period. [translated from Dutch]”


How Airwars assessment indicates Defensie is wrong

A review by Airwars shows that based on official Coalition data, the casualty spike in the NDU graphic can in fact only be explained if the Hawijah event had been included – indicating that US officials have long privately counted those 70 deaths in their own data, despite the Netherlands hiding its own involvement in the event.

For its assessment, Airwars examined all confirmed (‘Credible’) Coalition civilian harm events declared for the time window of May 1st to July 31st 2015. There were 17 such events totalling 106 confirmed deaths and 9 injuries if Hawijah was included – or 16 events with 36 deaths and 9 injuries if Hawijah was excluded, as Defensie claimed was the case.

However, the NDU graphic makes clear that the 2015 casualty spike relates only to Iraq – meaning that twelve Syrian events should be discounted. Three of the incidents had also been confirmed only in 2019, meaning that they were deemed Credible only after the NDU study was published.

That left just three events in Iraq: Hawihjah with 70 deaths; and two incidents in July 2015 each injuring one civilian according to the official data.

The casualty spike in the NDU graphic could only therefore be explained if the Hawijah estimate of 70 deaths had been included in the official tally, Airwars concluded.

“Defensie appears to have made a major error in claiming to the Dutch parliament that Hawijah  was excluded from the NDU study data,” notes Airwars director Chris Woods. “In fact, the only possible explanation for the visible casualty spike in summer 2015, depicted in the NDU graphic, was that the Hawijah event was already being included privately by the Pentagon in its own assessments of civilian harm resulting from Coalition actions.”

Screenshot of Airwars assessment of all declared Coalition civilian harm events for the period May 1st 2015 to July 31st 2015.

▲ The controversial NDU graphic which indicates that by 2018, the Pentagon was already privately counting a major loss of civilian life in Iraq during the summer of 2015, at the time of the Hawijah event.


April 2, 2017

Written by

Chris Woods

The US-led Coalition has conceded that a supposed ‘ISIS headquarters’ it targeted at Mosul in September 2015 was in fact a family home, noting in its latest civilian casualty release that “four civilians were unintentionally killed and two civilians were unintentionally injured in the building.”

Four members of the Rezzo family died when Coalition aircraft bombed their suburban Mosul villa on the night of September 20th-21st 2015. Despite a record 558 days between the incident and the Coalition’s public admission of error on April 1st, officials had known of possible civilian deaths within hours of the attack.

“This report was opened and a credibility assessment completed in 2015. However, the report was never officially closed or reported publicly. I do not know why that was,” Colonel Joe Scrocca, Director of Public Affairs for the Coalition told Airwars. “The case was brought to our attention by the media and we discovered the oversight, relooked [at] the case based on the information provided by the journalist and family, which confirmed the 2015 assessment, and officially closed the report in February.”

There was relief among family members that the deaths had finally been admitted – but also concern: “For eighteen months, we have been fighting for this admission of a mistake, for our loved ones to be counted as civilians,” Professor Zareena Grewal told Airwars from New York. “It is a small relief to have the US government concede that this airstrike was a mistake, that they mistakenly targeted the residential homes of a family that opposed ISIS. It is also deeply frightening because this case is an indictment of the quality of US intelligence.”

The Coalition admission – one of five newly confirmed civilian casualty events, all in Mosul – brings to 229 the number of Iraqi and Syrian civilians so far admitted killed in the US-led air war against so called Islamic State (ISIL or ISIS.) Airwars presently estimates that at least 2,831 civilians have so far died as a result of Coalition actions.

A family’s home destroyed

Among the declared targets struck by the US-led alliance on September 20th 2015 were “an ISIL VBIED facility, an ISIL bunker, an ISIL building, [and] an ISIL C2 node.” Now the Coalition says it also conducted “a strike on what was evaluated at the time to be an ISIS headquarters building.”

Cousins Najib and Tuka, both killed in a Coalition airstrike on September 20th-21st 2015 (Picture courtesy of the Altalib family)

Instead the home of a middle class family was destroyed. University professor Mohannad Rezzo; his 17-year old son Najib Mohannad Rezzo; his brother Bassim’s wife Miyada Rezzo and their 21-year old daughter Tuka Rezzo all died.

“Mohannad’s wife, Sana, survived the explosion, which flung her, burned, from her second-floor bedroom to the driveway below. Mohannad’s older brother, Bassim, also narrowly survived,” US-based relative Zareena Grewal wrote in the New York Times just days after the strike. “Bassim’s pelvis and leg were shattered in the attack and require surgery, but it is his emotional pain that consumes him.”

According to CENTCOM, military officials were aware of civilian casualty allegations within a day of the incident. Professor Grewal noted on October 4th 2015 that she had already been told that “Centcom was assessing the credibility of the reports, before determining any follow-on action, which might include a ‘formal investigation.'”

Yet despite Rezzo family members long ago coming forward with key photographic and other evidence, the alliance has continued publicly to deny any casualties until now. So confident were officials they had destroyed the right target that for more than a year, an official video of the Mosul attack was posted on the Coalition’s YouTube channel. It has since been removed, though not before being preserved by a pair of reporters who have been instrumental in helping secure a public admission of the Coalition’s error.

The Coalition’s own video of its attack on the Rezzo family home – since removed from its official YouTube channel

‘A long time coming’

Investigative journalists Azmat Khan and Anand Gopal have spent more than a year working closely with family members to secure an admission from the Coalition that it made a deadly error.

“Today’s official recognition of this airstrike having killed civilians has been a long time coming, and should have been made public previously. It is also a searing reminder of the immense difficulty families face in getting the loss of their loved ones recognized, even in cases in which there is ample evidence of civilian loss,” Azmat Khan told Airwars in an emailed comment.

“There is still information that the Coalition has refused to provide us, for example, the kind of aircraft and munitions used in this airstrike, as well as the reason why the Rezzo family homes were hit. We are also still awaiting the results of our Freedom of Information Act requests for the government’s own investigations into this incident.” Khan and Gopal’s major investigation into the incident is expected to publish in the near future.

Family members – while welcoming the official admission that their relatives were accidentally slain – remain angry that the process took so long. “Despite eyewitness testimony, a UN investigation, photographic evidence, and video footage of the strike that clearly demonstrated Coalition forces had hit two residential homes, the Pentagon did not count our family members as civilian victims and simply lumped them together with the death toll of Islamic State fighters,” says Professor Grewal. “The claim that our military air strike campaigns are precise is a dangerous and bloody myth.”

“We regret the unintentional loss of civilian lives resulting from Coalition efforts to defeat ISIS in Iraq and Syria and express our deepest sympathies to the families and others affected by these strikes,” the Coalition noted in its own press release.

Asked how the Coalition could have mistaken a family home for an ‘ISIS headquarters, a spokesman told Airwars: “ISIS uses many different types of structures to plan its terrorist activities. Many of which are residential homes taken from the people of Iraq and Syria.”

Backlog of allegations

The Rezzo admission is one of five new Mosul cases confirmed by the Coalition in its latest monthly civilian casualty report.

A Coalition strike on ‘an ISIS weapons manufacturing facility’ on January 30th 2017 is now thought to have unintentionallly killed one civilian in the building according to officials. Airwars understands that this event took place at Tanak neighbourhood, where up to 11 civilian deaths were reported by ISIL in a Coalition attack that day. Among those said by local monitors to have been slain were a young man Mustafa Mayser Mahmoud, his mother, and his father Mayser Mahmoud.

On February 6th the Coalition now says that “during a strike on ISIS fighters, it was assessed that three civilians were unintentionally injured when they entered the target area after the munition was released.” A similar attack against an ISIL truck bomb facility six days later also saw two civilians accidentally killed “when they entered the target area after the munition was released.”

The previously-unknown fifth incident on February 16th, again on “an ISIS VBIED facility” – this time in West Mosul’s Ar Rabi neighbourhood – killed a further two civilians according to officials.

Airwars is currently seeking to ascertain whether all five newly confirmed events were, as on previous occasions, the result of US-only actions.

In a mark of how steeply civilian casualty allegations are now rising, the Coalition announced in its latest report that it is still assessing 36 additional claimed civilian casualty events for February – on top of six more incidents for the month it has already deemed ‘not credible.’ Even so, this record monthly tally of 45 events under investigation still represents only half of the 90 claimed cases for February so far tracked by Airwars.

The international alliance admits it is falling behind on claims, though insists it intends to work through all cases: “The Coalition does have a backlog of allegations it is currently waiting to assess, to include additional allegations brought to our attention by Airwars. Credibility assessments take time and manpower to complete thoroughly,” Colonel Scrocca said in an emailed statement.

“While the primary mission of the Coalition is to defeat ISIS in Iraq and Syria, we should not and will not rush through this process for the sake of expediency. We take this responsibility very seriously and will continue to scrupulously assess every single allegation to ensure a full accounting of our findings.”

Mustafer Mayser Mahmoud died with his father (right) and mother in a reported airstrike on January 30th 2017, which the Coalition now appears to have conceded killed at least one civilian (via Mosul Ateka)