Nieman Foundation at Harvard
HOME
          
LATEST STORY
Israel has posted 1,200+ videos of airstrikes. This visual investigation shows the view — and the costs — on the ground in Gaza.
ABOUT                    SUBSCRIBE
Oct. 22, 2024, 12:37 p.m.
Reporting & Production

Israel has posted 1,200+ videos of airstrikes. This visual investigation shows the view — and the costs — on the ground in Gaza.

“By revealing the impact of these strikes, we were able to use footage released for one purpose to show a different perspective on the incidents.”

Over the last year, while the Israel Defense Forces posted footage on social media of its airstrikes claiming to target Hamas and other militant groups, Palestinians in Gaza have used platforms like Instagram and TikTok to document life under those airstrikes.

An investigation published earlier this month shows how airstrike footage and on-the-ground testimonies can be combined to advance war reporting. The investigation, by British news outlet Sky News and U.K.-based nonprofit Airwars, mapped and visualized civilian deaths, using the IDF’s social media posts and geolocation work by volunteers to find and corroborate information about the families and communities impacted by each strike.

Airwars and Sky News looked at IDF’s posts of 1,219 airstrikes in Gaza between October 7, 2023 and August 31, 2024. They geolocated 70 of those airstrikes, then matched coordinates of 17 to Airwars’ database of civilian harm. Those 17 airstrikes reportedly killed more than 400 Palestinians, the investigation found.

The two publications shared open source intelligence tools and worked with Sky News’ journalists in Gaza to corroborate and verify information. The outlets published separate stories about the findings; Airwars’ coverage, a website called “The Killings They Tweeted,” includes its first-ever film and an interactive map with links for each strike.

The project involved 10 reporters, three designers, and 16 geolocators, 14 of whom were volunteers. I chatted with Airwars’ head of investigation Joe Dyke and the project’s lead writer Rowena de Silva about conducting a social media-based investigation, getting around technical challenges, and designing a visual investigation with empathy for the viewer.

Our conversation below has been lightly edited for length and clarity.

Hanaa’ Tameez: How did this project come about?

Joe Dyke: Since the war in Gaza began, the Israeli military has posted an unprecedented number of clips of airstrikes on social media. These videos are typically grainy, black and white clips, often with five or more strikes spliced into a single video.

On their own, the clips don’t show you very much. They typically give no details on exactly what was the target and who was killed. The captions often say they are striking Hamas or other Palestinian militant groups, but there is usually no or little evidence. They never reference any civilians unintentionally harmed.

At Airwars, we work to question the rhetoric of militaries in their approach to civilian harm, and to uncover the reality of what is happening on the ground in conflict zones. So these clips were an immediate potential source of investigation for us [as we sought] to evaluate the Israeli military’s claims about these strikes against the reality on the ground.

We realized we were uniquely placed to conduct this investigation. For the past year, Airwars’ casualty recording team has been documenting allegations of civilian harm from airstrikes in Gaza. It’s resulted in hundreds of incident-by-incident reports of civilians publicly reported killed or injured — including exact geolocations of the incidents where possible. [We saw] a unique opportunity to try and match two versions of the same event — the sanitized footage of the strike from above, and the documented human impact on the ground.

Ultimately, we geolocated more than 70 strikes that the Israeli military published footage of. In 17 of these incidents, Airwars was able to match the footage to the exact geolocation of a documented civilian harm incident. In these 17 strikes alone, more than 400 civilians were reportedly killed.

Tameez: How did Sky News and Airwars link up? What was the collaboration like?

Dyke: We’ve talking to the [Data and Forensics] team at Sky News throughout the course of the war in Gaza. This team has published some world-leading investigations and vital reports about the conflict that has informed our own work and investigations. So we were keen to work with Sky’s investigations team.

Once we had a large enough sample size of incidents where we had matched the IDF footage to the civilian harm incident, we took the idea to Sky. Sky’s investigation team had done some similar work and had a number of cases that weren’t in Airwars’ published dataset, including two incidents in recent months.

Airwars’ investigations are primarily conducted using open-source techniques, which can create a distance between the evidence presented and the reality of violence experienced by civilians. But our aim is always to humanize civilians who are harmed by conflict. In addition to their significant support on the story and narrative, collaborating with Sky gave us a unique opportunity to access information from their Gaza team reporting on the ground. This included access to images, videos, and testimony accounts from witnesses, and ultimately it allowed us to better center those perspectives in our work.

Tameez: Why is this investigation important to publish now?

Rowena De Silva: The information environment around the war in Gaza is complex, with misinformation and disinformation circulating quickly, especially on social media. With international journalists prevented from entering Gaza except on organized tours with the Israeli media, OSINT investigations have become more vital in verifying the impact of these strikes on civilians. Palestinians are documenting their experiences of war — listing the deaths of their loved ones on social media and publicly reporting the impact of strikes across the strip. Airwars seeks to preserve this information by providing an archive of civilian testimonies to better understand the human toll of conflict.

The Killings They Tweeted is one part of this larger organizational effort to document civilian harm in Gaza. In analyzing the strike footage published by the IDF, along with claims of exclusively striking Hamas infrastructure or targeting Hamas militants, we were able to uncover more of the truth behind such military statements. By cross-checking strike footage with Airwars’ own civilian harm archive, we revealed a more complete picture of civilian harm, one that differed from the narrative of precision that the IDF continues to push in its military campaigns.

Tameez: Airwars and Sky News have published OSINT investigations before, but what stands out to you about this project? How and why does it differ from previous investigations?

De Silva: In terms of content, the power of this investigation comes from the fact that it’s interrogating strike footage tweeted by the Israeli military itself. It stands out from our other pieces because the material under review was publicly released by the military to portray a positive image of its actions. By revealing the impact of these strikes, we were able to use footage released for one purpose to show a different perspective on the incidents.

This is also the first short film Airwars has produced in-house, and one of the first Airwars investigations that relies so heavily on the work accumulated by the geolocation team.

Tameez: How did the collaboration with volunteer geolocators work?

De Silva: There’s a dedicated group of geolocation volunteers at Airwars. Many of them decided to join the team in the months after the war in Gaza began. We’ve seen a large amount of images and videos uploaded by Palestinians on social media in Gaza, testifying to their experiences of airstrikes. The geolocation team has been working on locating every one of the hundreds of allegations of civilian harm in Gaza researched by the casualty recording team at Airwars. This provided an extensive database for us to cross-check the IDF footage with civilian harm incidents.

This primary process — of geolocating civilian harm incidents — was where the geolocation volunteers were primarily used. They weren’t involved in the day-to-day process of seeking to match the IDF footage with these incidents, or in the production of the film.

Tameez: How did you all get around the blurriness of the IDF videos? What are some of the other challenges that came up in the reporting and fact-checking, and how did you address them?

Dyke: We began by reviewing more than 600 clips of strikes released by the Israeli military in the first month of the war and assessing the “geolocatability” of each one. We implmented a grade system to indicate the likelihood of finding the exact location of the strike in the videos.

We assigned each strike clip a number between 1 and 5 based on how clear the imagery was, if the strike hit a distinctive building such as a mosque or school, or if there were any unique urban features that would make it more noticeable in satellite imagery. We then prioritized geolocating the “high-scoring” videos. If a strike clip was incredibly blurry and showed a nondescript building being hit, it would likely have been deemed “low-scoring,” making it less of a priority for potential geolocation. Once this was completed, there were several methods of matching “high-scoring” IDF strike footage with civilian harm incidents.

If the strike hit civilian infrastructure, we cross-checked the Airwars archive. For example, if a video of a strike on a mosque was published on October 17, we reviewed all strikes on mosques documented by the casualty recording team on October 17. We geolocated the mosques listed in the internal Airwars research documents and compared satellite imagery with the IDF footage to see if there was a match. Early on in the process, we realized the IDF [was also publishing] footage of strikes that occurred several days prior, so we expanded the date range when cross-checking civic buildings with the Airwars archive.

We compared every geolocated incident of civilian harm completed by the Airwars geolocation team with “high-scoring” IDF strike footage. For example, if a strike was geolocated to a residential building by Airwars on October 17, we compared the satellite imagery with clips of strikes published by the IDF on and around October 17. This was the most direct method of linking published IDF footage with cases of civilian harm.

In other instances, we geolocated “high-scoring” IDF footage by manually searching across satellite imagery of Gaza. There was little or no geographical information provided alongside the IDF strike footage. By analyzing the density of the area pictured in the clip, the investigations team assessed whether the strike was in a rural area or city environment, narrowing down areas of Gaza to parse through. By matching distinct features with satellite imagery, we identified exact locations of IDF strikes.

Overall, we geolocated 72 IDF strikes and matched 17 to civilian harm incidents. While 55 clips of IDF strike footage remain unmatched to the Airwars archive, that doesn’t definitively mean no civilian harm occurred in these incidents. Airwars documents public allegations of civilian harm, but many civilian deaths aren’t publicly reported on and therefore wouldn’t appear in the Airwars archive. In addition, if the Airwars geolocation team had limited evidence on the exact location of a strike that caused civilian harm, it would be impossible to conclusively match an IDF clip of a strike to that incident.

Tameez: How did you decide on a map, short film, and text as the main elements? Why did you blur images and require users to click to view them?

De Silva: Our aim was to give the audience as much understanding of the nature and volume of the IDF strike videos as possible. There are hundreds of clips, and we wanted them to see firsthand what these videos look like and the information the military provided in the captions. This allowed us to clearly juxtapose the black and white, sterile military strike footage taken from above with the destruction captured in videos taken by civilians on the ground. By spotlighting three strike clips where civilians had been killed, we delved into larger questions around Israeli military practice and the disproportionate harm to civilians in Gaza.

The interactive map was a method to visualize our findings. We wanted to add as many layers of context to the geolocated strike footage as possible to combat the clinical presentation of the strikes posted by the Israeli military. Where possible, we listed the number of civilians allegedly killed, along with as much detail on the location and date of the strike as possible.

Each video matched with civilian harm contains a link to the relevant report in the Airwars archive. The research gathered by the casualty recording team is preserved here, including testimonies and witness accounts, the names and pictures of civilians killed in the strike, details on who they were, and videos and images connected to the incident. We believe transparency and replicability is vital in visual and OSINT journalism. In theory, any researcher using the footage and the coordinates we have published could replicate the process and test the findings of our work.

The media attached to civilian harm incidents in our Gaza archive is automatically blurred. Palestinians in Gaza are posting substantial imagery on social media. That provides an essential level of verification to witness statements, but much of the imagery is highly graphic. We made an organization-wide decision to blur it, allowing the viewer to choose to remove the filter for each image and consciously view these images of violence if they choose to do so.

Tameez: What has the response to this project been like so far? What’s next?

Dyke: The combination of the strong visual identity of the investigation paired with the investigation’s extensive open source methodology and powerful findings has created traction in the civilian harm field as well as media and legal circles. It has been widely shared across social media, with overwhelmingly positive feedback. It was shared by many of the most well-known people in the OSINT investigative industry, including the founder of Bellingcat and the head of the NYT Visual Investigations team.

We’ll continue to push the articles and film in various forums over the coming months, with a focus on routes for accountability.

Palestinians inspect the ruins of Aklouk Tower destroyed in Israeli airstrikes in Gaza City on October 8, 2023. Photo by Palestinian News & Information Agency (Wafa) in contract with APAimages.

Hanaa' Tameez is a staff writer at Nieman Lab. You can reach her via email (hanaa@niemanlab.org), Twitter DM (@HanaaTameez), or on Signal (@hanaatameez.01).
POSTED     Oct. 22, 2024, 12:37 p.m.
SEE MORE ON Reporting & Production
Show tags
 
Join the 60,000 who get the freshest future-of-journalism news in our daily email.
Don’t trust the polls? Neither did The New York Times in 1956 (spoiler: it didn’t work out great)
“There was some feeling that the Times should stick to reporting trends and let the pollsters make the forecasts.”
A year in, The Guardian’s European edition contributes 15% of the publisher’s pageviews
After the launch of Guardian Europe, one-time donations from European readers increased by 45%.
Press Forward awards $20 million to 205 small local newsrooms
In response to the volume and quality of applications, Press Forward doubled the funding and number of grantees for this open call.