Proving Patterns of
Cruelty from Afar

How researchers exposed the abusive military strategy behind attacks on civilians in northern Syria – without visiting the sites

Civilians in Syria’s war zones are on their own these days, with different sides doing everything they can to stop independent witnesses, like journalists and human rights researchers, from reaching the battlefield. Despite this, in October 2020, Human Rights Watch published a 167-page report on 46 attacks by Syrian and Russian forces that destroyed civilian infrastructure in Idlib governorate, Syria. Without setting foot on Syrian soil, we exposed apparent war crimes and the abusive military strategy that drove more than one million civilians from their homes. We wanted to show you more about the methods and approach.

Airstrikes in Idlib

Belkis Wille, Senior Crisis and Conflict Researcher. Based in Yangon, Myanmar

By March 2020, Idlib was one of the last anti-government holdouts in northwest Syria. Fighting had raged in Syria for almost a decade, creating 5.5 million refugees and 6.2 million internally displaced people, and rebels threatened the autocratic rule of President Bashar al-Assad, until he forged a military alliance with Moscow. Since then, a Syrian-Russian military alliance repeatedly bombed hospitals, schools, and markets. The United Nations, foreign governments and human rights groups condemned these attacks, which violated the laws of war, to no avail.

We were so frustrated about documenting the endless atrocities with limited effect. We wanted to look further, to dig deeper and examine the strategy used by the Syrian and Russian militaries that caused so much harm to civilians. We felt it wasn’t a coincidence they kept bombing hospitals and schools, the lifeline for civilians still left in Idlib. So we examined the pattern of attacks, to test a common allegation: that Syria and Russia were intentionally targeting civilian infrastructure to drive Idlib’s civilians from their homes and help the Syrian army to retake territory. Understanding the military strategy behind the attacks proved critical for us to then craft the right advocacy messages to those in power, to get them to bring a stop to war crimes in Syria.

Syria's President Bashar al-Assad (left), Russia's President Vladimir Putin (right), Minister of Defense Sergei Shoigu (second right), and Lieutenant-General Alexander Yuryevich Chaiko (third right) during a meeting.

Syria's President Bashar al-Assad (left), Russia's President Vladimir Putin (right), Minister of Defense Sergei Shoigu (second right), and Lieutenant-General Alexander Yuryevich Chaiko (third right) during a meeting. © Alexey DRUZHININ / SPUTNIK / AFP

President Putin makes remarks at the Damascus command post for Russian forces during a January 7, 2020 trip to Syria.

Screenshots from video show President Putin making remarks at the Damascus command post for Russian forces during a January 7, 2020 trip to Syria. Russian Federation Ministry of Defense

In early March I flew to Antakya, southern Turkey, just across the border from Idlib, where a ceasefire had just been agreed. The Turkish government, which keeps the border closed, didn’t grant me permission to cross, so I met people who had escaped Idlib, people who were in contact with rescue workers, doctors, teachers, and lawyers still trapped in the besieged region. I wasn’t sure if we could gather enough evidence to make our case without entering Syria.

At a café in Antakya one evening, a Syrian lawyer told me about an attack he witnessed in Idlib City on January 15. A strike had hit a grocery stall in the vegetable section of a main market, he said, just after 2 p.m. His friends put us in touch by phone with another witness to the attack, Ayman Assad, who ran a car repair shop nearby.

Ayman Assad stands in front of his shop in Idlib city, northern Syria.© Ayman Assad

Ayman said that a few minutes after he heard the explosion in the market, two bombs hit shops just 50 meters from where he was standing. Ayman and the lawyer, as well as rescue workers we interviewed, said the attacks had a devastating impact. From their accounts, we determined that the attacks killed at least 22 civilians, wounded another 65, and destroyed at least 20 shops. The entire market was shut for several days.

I called my colleague Richard Weir to see if his Syrian contacts could help us investigate.

Richard Weir, Crisis and Conflict Researcher. Based in Beirut, Lebanon

I knew when Belkis called that we had to find a way to make sense of each incident in this stream of attacks and that we would need a variety of sources to help us understand them. In this case, we had an incident, date, time, location, and a witness, but we needed more.

In Idlib, a network of people monitor aircraft movements in Syria and share encrypted messages on the presence of aircraft, sometimes identifying the type of aircraft they are seeing from below. I opened the Telegram messaging app, scrolled through two months of messages, and found posts from the afternoon of January 15, 2020. The group’s messages said the network had identified a plane flying over Idlib City moments before the two attacks. It seemed likely that this plane could have launched the strikes that Ayman Assad and others said they witnessed that day.

Backing Up the Testimony

Gabriela Ivens, Head of Open Source Research. Based in Berlin, Germany

We had a date, general location, and reports of a plane overhead at the time. We need to find the precise coordinates for both attacks, verify evidence of damage, and look for any signs of military targets that might justify an attack. We asked the archival and investigative project Syrian Archive to send us all videos posted on YouTube between October 2019 and February 2020 that included certain keywords, like Syria, Syrian conflict, Idlib, and so on.

By going through their spreadsheet of videos, using a mix of keyword searches and advanced location search techniques in Arabic and English, I found dozens of videos of both attacks, all posted on social media on January 15 and 16 by people who had recorded the aftermath of the attacks. Some videos captured both attacks.

Top left © Hadi Al Abdallah. Bottom center © Baladi-News Network. Bottom right © Orient News. All other videos © Syrian Civil Defense

One video showed two people who had been killed lying amongst vegetables. Behind them, something caught my eye: glimpses of a distinct white building, a large structure with unique panels on the windows. I could see a fence separating this building from the road where the vegetable market was attacked.

This building seemed like my best chance to find the exact coordinates. From the interviews Belkis had done, I understood the vegetable market was in the northeast of Idlib City. Looking on Google Maps for possible matches of this building, I came across Idlib Museum, a large white structure.

Idlib Museum on August 25, 2018. Omar Haj Kadour / AFP

Idlib Museum on August 25, 2018. © Omar Haj Kadour / AFP

Those photos only showed the front of the museum, which wasn’t a match to the scene in the video. I then found a photograph posted by someone who visited the museum years ago that showed the back of the building. It matched perfectly.

Google Images

Screenshot from Google Maps showing geotagged images near Idlib Museum.

I saw the same detail on the windows, and the fence around the museum. Cross-checking the location of Idlib Museum with satellite imagery, I could see the stalls in the vegetable market right beside the building.

Video © Baladi-News Network. Satellite image dated October 27, 2018 © Maxar Technologies. Source: Google Earth. Photograph: © Hashem al-Abdullah

I knew the second blast site had to be nearby – the witness, Ayman, said he had heard the market attack from his shop. I went through more online videos.

One, taken by rescue workers, contained drone footage that showed a fire-engine spraying water over shops and stalls that had been reduced to rubble, matching what Ayman and the lawyer had described.

© Syrian Civil Defense

I was able to map out the shops and roads in the area by comparing the drone footage to satellite images of the area. Now I had identified both attack locations.

Verifying the Destruction

Carolina Jordá Álvarez, Geospatial Analyst. Based in Geneva, Switzerland

My colleagues presented me with all the evidence of this attack gathered so far: witness testimony, flight spotter logs, and videos we had verified as being taken at the time and location of the attacks, which allowed us to figure out the exact coordinates where the bomb blasts had occurred. Now I needed to assess the scale of the damage.


Searching through commercial satellite imagery, I found a high-resolution satellite image of the location taken two months after the attack.

I could see that a blast had destroyed several markets stalls and had impacted the eastern part of the fence around the museum.

The impact crater, only five meters from the market, was still visible months after the attack.

Tracing my fingers down my screen, less than 300 meters south of the first attack, I could see another impact site where several industrial warehouses had been hit multiple times.

Satellite image dated March 4, 2020 © Maxar Technologies. Source: EUSI

Assessing the damage from the high-resolution image, I went to another satellite service that captures areas of the world more often but at a lower resolution.I found an image taken a day before the attack that showed no damage in either of these locations, and an image from afterwards – it made the damage from the attack very clear.

Satellite images © Planet Labs

Searching other satellite imagery platforms, we could also see damage to the industrial area as described by those we interviewed.

Belkis Wille, Senior Crisis and Conflict Researcher. Based in Yangon, Myanmar

These are just two examples of how we pieced together evidence of attacks. Sometimes our research started with the words of a witness. Other times we started with online videos of an attack we had not yet heard about, then went back to our sources to find witnesses. And in some cases, analyzing satellite images led us to see patterns and discover facts we might have missed. We went through these steps many times, verifying another 44 ground attacks and airstrikes between March 2019 and March 2020 that killed at least 202 more civilians – in hospitals, schools, markets, homes and camps for displaced people. In none of these attacks did we find any presence of anti-government armed groups. These looked like war crimes, and a pattern of attacks on the civilian population that could amount to crimes against humanity. We needed to understand the strategy behind such attacks.


Areas of control data: Syrian Network of Human Rights

Russia’s Role and Military Strategy

Richard Weir, Crisis and Conflict Researcher. Based in Beirut, Lebanon

Going back to the flight spotter channels, we used information, including the specific aircraft cited, the airbases used for take-off, and some intercepted radio communications, to identify specific planes used in the attacks. This helped us to attribute six such attacks to the Russian armed forces.

Searching through social media, online documents, official media, and information gathered from interviews with experts, I mapped out the Russian military structure in Syria and the role of the Russian armed forces at the highest levels in the shaping and implementation of military operations in Syria. I incorporated the names and ranks of three Russian commanders sent to Syria, allowing us to build a chain of command and a list of Russian individuals responsible for the Idlib operation as a whole, and for the specific attacks described in our report. I found different parts of the puzzle by reading updates from Russia’s Ministry of Defense, posts about senior Russian military officers killed in combat in Syria, and interviews with senior officials proclaiming the success of their military strategy. I examined the captions, names, and titles of individuals photographed during briefings about military operations in Syria and noted who was awarded medals for service there. While this information was publicly available, it had not been been pulled together in one place.

I also looked at the impact of the attacks on people’s movements. Using statistics provided by humanitarian organizations, we found that in multiple areas, people abandoned their homes en masse following strikes on hospitals, schools, and other civilian infrastructure.

Syrian families displaced by attacks carried out by the Syrian-Russian military alliance made their way to areas less under attack near the Turkish border in Idlib, northwest Syria in February 2020. © Ibrahim Dervis/Anadolu Agency via Getty Images

LEFT: Syrian families displaced by attacks carried out by the Syrian-Russian military alliance made their way to areas less under attack near the Turkish border in Idlib governorate, northwest Syria in February 2020. RIGHT: A Syrian child looks through a tent during hot weather at a camp where displaced Syrians shelter in Idlib governorate, northwest Syria in September 2020. © Muhammed Said/Anadolu Agency via Getty Images

Entire villages were often left abandoned within days of the strikes. Available data show more than 200 villages were emptied – and left undefended by anti-government armed groups. Syrian ground forces, severely weakened from years of fighting, were able to retake control of these villages virtually unopposed.

Abandoned buildings in Ariha town in Idlib governorate, northwest Syria on February 19, 2020. © Muhammed Said/Anadolu Agency via Getty Images

This research showed that the targeting of civilian infrastructure was more than coincidence or a series of errors. The Syrians and Russians continued to claim that they did not target civilians, only attacking those they considered to be “terrorists” – despite the pattern of attacks that appeared to exclusively harm civilians. The Syrians even sent a letter to the United Nations Security Council in mid-2019, attempting to justify some attacks by falsely claiming that over 100 hospitals had been taken over by “terrorists” and that ambulances were compromised. This overly broad use of the “terrorist” label and repeated attacks that progressively crippled communities and forced people out suggested a broader, thinly veiled strategy: to weaken support for the anti-government armed groups in part by imposing suffering on civilians. Academics call this punishing the population.

The response from Russian authorities is to deny any laws-of-war violations in Idlib, despite the wealth of evidence to the contrary. This made it all the more important for us to get as much detail as possible on the exact dates, times, locations, impact, and stories around each attack to counter Russia’s denials. In one case, the Russian military tried to claim an alleged market attack didn’t happen by showing a photo of a completely different market, still intact.

Rebutting Russia’s Denials

Gabriela Ivens, Head of Open Source Research. Based in Berlin, Germany

On March 2, 2020, a report released by the UN Independent International Commission of Inquiry on Syria included the following account:

[F]light spotters had reported that at least two Russian fixed wing-aircrafts departed from Hmemim airbase on July 22 at around 8:03 a.m. and circled near the target area at the same time of the attack. The Commission further obtained flight communication intercepts conducted in the Russian language for the period between 07:40 and 09:59 a.m. The communications confirm that, between 8:17 and 8:35 a.m., two Russian aircraft operated in Idlib airspace, and reported on the progress of on-going aerial attacks in Idlib.

Videos and photographs posted online on July 22, 2019, showed a large explosion that turned several buildings in a marketplace into rubble and left numerous bodies, some torn to pieces.

© Hadi Al Abdallah

Searching online for more information, I found a Russian officer who spoke and presented photographs at a July 29 news conference. He claimed that after the “alleged” attack Russian planes had flown over and confirmed the marketplace had never been hit and was functioning as normal.

Russian Federation Ministry of Defense

Carolina Jordá Álvarez, Geospatial Analyst. Based in Geneva, Switzerland

This Russian claim did not match what witnesses of the attack told us. By analyzing lower-resolution satellite imagery from the area, I found two images taken at 9:51 a.m. Damascus time on July 21 and at 9:51 a.m. on July 22, and I got confirmation: it was clear from these two images that the attack took place during this 24-hour window.

Satellite images © Planet Labs

Accessing higher-resolution imagery allowed me to assess the area in more detail, including the number of buildings affected.


I realized what the Russian military had done at the press conference: the photographs of the market claimed as “intact” in fact showed a totally different place –

a wholesale food market about 350 meters away from the site attacked on July 22.

Satellite image dated September 23, 2019 © Maxar Technologies. Source: Google Earth

The wholesale market was itself attacked five months later, on December 2. This low-tech misdirection wasn’t anything new. My colleagues documented a similar denial by the Russian government in 2016, when the same officer used imagery to mislead the public about an attack on a hospital.

Identifying the Weapons

Belkis Wille, Senior Crisis and Conflict Researcher. Based in Yangon, Myanmar

We weren’t only relying on social media for visual evidence of the attacks. Rescue workers and witnesses also sent us photographs and videos they had taken during and after the attacks. On the morning of March 4, 2020, students and teachers at Walid Shabaan secondary school in the town of Jisr al-Shughour found an unexploded munition in the schoolyard and called the Syrian Civil Defense (also known as the White Helmets) to remove it.

The SCD’s ordnance disposal team found the remnants of a cluster munition rocket and unexploded submunition in the yard of Walid Shaaban secondary school in Jisr al-Shughour after an attack on March 3, 2020. © Syrian Civil Defense

The Syrian Civil Defense team leader told us over the phone that on the previous night there had been heavy shelling and airstrikes, that cluster munitions had landed near the school building, and that one submunition landed in the yard.

Mark Hiznay, Arms Division Associate Director. Based in Washington, DC, USA

Belkis sent me a series of photos showing remnants of ground-launched Russian-made cluster munition rockets found after attacks on Idlib schools in February and March. Identifying them was not difficult given the images I received, which I was able to compare with images of remnants I had taken years ago in Ukraine, identifying them was not difficult: empty cargo sections and unexploded 9N235 submunitions of 9M27K Uragan rockets and rocket motors of 9M55K Smerch rockets at three schools in the governorate.

These cluster munitions rockets are fired from multi-barrel launchers from more than 20 kilometers away and can saturate an area with hundreds of metal fragments the size of pistol bullets from the dozens of submunitions that each rocket can indiscriminately deliver.

A BM-27 Uragan MLRS, used as a monument to A.N. Ganichev near Splav State Research and Production Enterprise in Tula, Russia. © Vitaly Kuzmin

The unexploded submunitions in question had failed twice: first failing to detonate on impact with the ground as designed, and second after the back-up self-destruct system failed. No dud should be touched because they are sensitive to movement and must be destroyed in place by deminers. The photographs I reviewed from Belkis showed the Syrian Civil Defense team doing exactly that.

© Syrian Civil Defense

Identification of the cluster munition rockets from these attacks was relatively simple because we have seen the same weapons used in Syria since 2014. We also found them in Ukraine in 2014-2015 and Georgia in 2008. To ease identification and reach other sources who may have witnessed attacks with the same weapons, we created and distributed graphics of what the remnants look like in the field for the Uragan and Smerch rockets as well as the 9N235 fragmentation submunition and encouraged people to share photos of other remnants they discovered.

Severe Damage

Richard Weir, Crisis and Conflict Researcher. Based in Beirut, Lebanon

Most of the 46 attacks we documented were carried out using explosive weapons with wide-area effects. We have long advocated for states to implement a strong political declaration against using these weapons in populated areas because of their indiscriminate and devastating impact on civilians. Knowing what weapons the Syrian-Russian alliance was using, and the impact they were designed to have, was also key to understanding the military strategy and its use of weapons to “punish” the population.

Aside from the damage and deaths caused by these strikes, indiscriminate attacks traumatize and terrorize people.

The Final Result

Belkis Wille, Senior Crisis and Conflict Researcher. Based in Yangon, Myanmar

When we started this project six months ago, we weren’t sure we could pull it off, given that it was impossible to do research on the ground in Syria. How could we document airstrikes and destruction without seeing it ourselves? We used a range of technology and remote research methods – reviewing dozens of satellite images, verifying more than 550 photos and videos, investigating the Russian and Syrian military command structures – combined with 113 traditional human rights interviews, albeit from a distance. That’s how we documented the patterns of attacks and painted a damning picture of the motivations behind the war crimes committed by Syrian and Russian forces.

As always, our research is the means to an end: to protect civilians in conflict and press for justice at the highest levels.

Read the report »
More on Human Rights Watch’s work on Syria »

We thank the individuals who made this research possible by sharing their experiences, as well as the Sentry and Syrian Civil Defense volunteers, whose courageous work was vital to identifying and investigating individual strikes. Thanks also to the civilians and journalists who recorded and uploaded videos and photographs of the aftermath of attacks. This documentation, often recorded at personal risk, enabled Human Rights Watch researchers to report on incidents with a higher level of accuracy and nuance. The archival and investigation group Syrian Archive also contributed greatly by providing details of over 4,000 videos that it had collected from social media platforms. Many thanks to Planet for providing satellite imagery that was essential for this investigation.

We would also like to thank our colleagues at Human Rights Watch who provided research support, reviewed, and edited the report and to our colleagues who produced the graphics and videos that brought this research to life.

Header video: Top row, second video © 2020 Ahmad Rahal. Middle row, center video © 2019 Hadi Al Abdallah. Bottom row, third video © 2019 Orient News. Bottom row, fourth video © 2019 Hadi Al Abdallah. All other videos: © 2019 Syrian Civil Defense. Satellite imagery © 2020 Maxar Technologies. Sources: EUSI and Google Earth.