by Azmat Khan and Anand Gopal, the New York Times
A key part of the coalition’s investigation process is to match civilian casualty accusations against its own logs. Chris Umphres, an Air Force captain at Udeid who assesses allegations of civilian casualties, told us that military investigators possess the coordinates of “every single strike conducted by coalition forces,” crucial information unavailable to the typical journalist. “We have 100 percent accountability of where all of our weapons are employed.”
We found this to not always be the case. For every location we visited, we submitted GPS coordinates to determine whether it was the coalition or the Iraqi Air Force that bombed the site. At first, the coalition told us it did not have the time or the staff to check more than a handful of the coordinates. But eventually, a team of Air Force analysts at Udeid agreed to compare the dates and coordinates of each of the 103 sites in our sample with those the coalition had recorded in its airstrike log. If a strike in our sample occurred within 50 meters of a strike that was recorded in the logs, they classified it as a “probable coalition airstrike,” while assessing those outside this range — that is, anything more than a couple of house-lengths away — as “unlikely.”
By this measure, 30 of the 103 strike sites in the sample we submitted are probable coalition strikes. But other evidence suggests that the coalition was responsible for many more. Human rights organizations have repeatedly found discrepancies between the dates or locations of strikes and those recorded in the logs. In one instance, the coalition deemed an allegation regarding a strike in the Al-Thani neighborhood of Tabqa, Syria, on Dec. 20, 2016, as “not credible,” explaining that the nearest airstrike was more than a kilometer away. After Human Rights Watch dispatched researchers to the ground and discovered evidence to the contrary, the coalition acknowledged the strike as its own.
We found many such discrepancies. For instance, the Air Force analysts said it was unlikely that the coalition had struck Qaiyara’s water-sanitation facility because the logs recorded the nearest strike as 600 meters away, which would place it outside the compound entirely. Yet we discovered a video — uploaded by the coalition itself — showing a direct strike on that very facility. (When we asked Lt. Col. Damien Pickart, director of public affairs at Udeid, about this discrepancy, he said he could only report “what the strike log shows.”) Similarly, we were told that a strike we identified on Qaiyara’s main bridge was unlikely to be by the coalition, because the nearest strike was on a truck 150 meters away. We again found a coalition video showing a direct hit on the structure. Pickart explained the inconsistency by saying the coalition had conducted multiple strikes on various targets within an hourlong period, only one of which was included in the official log.
The most common justification the coalition gives when denying civilian casualty allegations is that it has no record of carrying out a strike at the time or area in question. If incomplete accounts like these are standard practice, it calls into question the coalition’s ability to determine whether any strike is its own. Still, even using the most conservative rubric and selecting only those 30 airstrikes the Air Force analysts classified as “probable” coalition airstrikes, we found at least 21 civilians had been killed in six strikes. Expanding to the 65 strikes that fell within 600 meters — for example, the strikes on the home of Inas Hamadi in Qaiyara and the electrical substation in Aden — pushed that figure to at least 54 killed in 15 strikes. No matter which threshold we used, though, the results from our sample were consistent: One of every five airstrikes killed a civilian.
To understand how radically different our assessment is from the coalition’s own, consider this: According to the coalition’s available data, 89 of its more than 14,000 airstrikes in Iraq have resulted in civilian deaths, or about one of every 157 strikes. The rate we found on the ground — one out of every five — is 31 times as high.
Read the rest of the report here.