Feedback
News

Study: US drone strikes more likely to kill civilians than US jet fire

A U.S. Predator drone armed with a missile sits at the military airport in Kandahar, Afghanistan, on June 13, 2010. Massoud Hossaini / AFP - Getty Images file

Drone strikes — billed by President Barack Obama as tactically surgical and less deadly to civilians than conventional air power — are 10 times more likely to cause innocent casualties than bombs or missiles unleashed from U.S. jets, according to a new study based on classified military documents.

The report’s author, Lawrence Lewis, a researcher at the federally funded Center for Naval Analyses who possesses a top-security clearance, dissected secret data on U.S. air attacks in Afghanistan from mid-2010 to mid-2011 — the peak of unmanned drone use during the war, executed under the command of former Gen. David Petraeus.

After reviewing casualty statistics from that 12-month span, Lewis published a related article in the military journal Prism, criticizing Pentagon leaders for giving drone operators “limited training” on how to minimize civilian harm. He also called for the Defense Department to investigate the impact on civilians inflicted by U.S. drone attacks in Pakistan. (On Wednesday, a U.S. drone strike killed at least 17 people in Pakistan’s volatile North Waziristan tribal region.)

“Already, criticism over U.S. clandestine (drone) operations is putting the administration on the defensive and growing louder as local populations, particularly in Pakistan, join in protesting the use of (drones),” Lewis wrote in the journal article, co-authored by Sarah Holewinski, who heads the private, Washington, D.C.,-based Center for Civilians in Conflict. “... The U.S. government should undertake a review of the potentially negative impact of (drone) strikes, both in counterterrorism efforts and with regard to civilian harm,” Lewis wrote.

The report comes at a time when the Pentagon has called for a nearly one-third increase in its drone fleet and as DOD leaders have expanded use of the remotely controlled weapon in “non-declared theaters of armed conflict" — such as Pakistan and Yemen.

Civilians inadvertently killed or wounded by drone strikes "haunt" Obama, the president acknowledged while speaking May 23 at the National Defense University in Washington. In an address aimed at redefining American counterterrorism measures, he vowed to narrow the parameters for the use of drones to kill terrorists overseas and to limit collateral casualties. 

Lewis' findings affirmed that many innocent people have been caught in the drone crossfire, or misidentified as terrorists. But because the researcher was reviewing and reporting on classified documents, he was unable to reveal exactly how many civilians died in drone attacks in Afghanistan in 2010 and 2011, Holewinski said. Lewis was unavailable for an interview Tuesday. 

“If this finding can be so contradictory to what we inherently believe about drones in Afghanistan, well then certainly the administration has an interest in knowing if this is also the case in a place like Pakistan,” Holewinski said. 

She also confirmed that classified data reviewed by Lewis showed U.S. drone strikes were 10 times more likely to harm civilians than were weapons fired from manned platforms — a numeric rate not contained in the duo's published article nor in an accompanying, declassified executive summary drafted by Lewis. That figure was first reported earlier Tuesday by The Guardian.

“Our concern is that not only is that a problem in Afghanistan, but if you look at Pakistan, Yemen, Somalia, wherever else these drones are being used, the administration hasn’t said anything, hasn't been transparent, hasn't released any of the civilian-protection protocols, or (described) any of the training on civilian-harm mitigation that the drone pilots get,” Holewinski said. “When you don’t have the right training — which gives you the right mindset, the right skills, the right tactics, and the right way to think about the civilian population — you’re more likely to get more civilian harm.” 

Pentagon officials declined to discuss details of the classified study. But Commander Bill Speaks, a DOD spokesman, said the efforts of military leaders to minimize civilian casualties involve "a broad range of tactics and procedures" and apply to both manned and unmanned aircraft. 

"To portray it as a choice between the two platforms grossly oversimplifies the issue," Speaks wrote in an email. "Civilian casualties caused by our forces in Afghanistan have steadily declined in recent years, a trend that continues even as our reliance on remotely piloted aircraft (RPA) has increased."

During 2012, airstrikes by the NATO-led security mission in Afghanistan were responsible for 2.65 percent of the overall civilian casualties, according to Speaks. 

"We continue to further minimize civilian casualties — lessons are continuously shared, procedures improved, incidents investigated, policies modified, training amended and tactics updated," Speaks wrote. "The mitigation of civilian casualties has direct command oversight and the highest level of priority." 

The Afghanistan casualty details reviewed by Lewis were assembled because American forces often returned to the scene of drone attacks to interview locals and elders about who died, who was injured and whether the targets were indeed plotting violence. That kind of cross-checking is far more challenging in Pakistan, Yemen and Somalia, Lewis wrote.

"While the U.S. has repeatedly stressed how (drone) strikes in Pakistan cause very few civilian deaths, this position runs counter to independent investigations," Lewis wrote.

He cited a March 11, 2011, U.S drone strike on a vehicle in Pakistan during which a follow-up drone blast reportedly killed rescuers who had converged on the scene, and noted that "several reports stated there were civilian casualties, ranging from two to five individuals." Lewis also mentioned a March 17, 2011, U.S. drone strike on a suspected militant compound in that same country — after which Pakistani authorities and news reports stated that the gathering was actually a tribal assembly of elders intended to settle a dispute at a nearby chromite mine. The reports of civilian casualties ranged from 13 to 44.

In his executive summary on his study, Lewis wrote that "inaccurate assessments of civilian harm" by American military and political leaders have "tarnished the U.S.' reputation." 

"We’ve heard from the administration that what comes first is avoiding civilian harm, that it's one of their top priorities," Holewinski said. "I really want to believe that — especially about my own country. The thing is, they have to back that up with some kind of proof. They can’t just say: 'Trust us.' "

More from NBC News Investigations:

Follow NBC News Investigations on Twitter and Facebook