retour article original
par Ryan DEVEREAUX
From 2011 to 2013, the most elite forces in the U.S. military, supported by the CIA and other elements of the intelligence community, set out to destroy the Taliban and al Qaeda forces that remained hidden among the soaring peaks and plunging valleys of the Hindu Kush, along Afghanistan’s northeastern border with Pakistan. Dubbed Operation Haymaker, the campaign has been described as a potential model for the future of American warfare : special operations units, partnered with embedded intelligence elements running a network of informants, pinpointing members of violent organizations, then drawing up plans to eliminate those targets from the battlefield, either by capturing or killing them.
The intelligence community documents obtained by The Intercept, detailing the purpose and achievements of the Haymaker campaign, indicate that the American forces involved in the operations had, at least on paper, all of the components they needed to succeed. After more than a decade of war in Afghanistan, a robust network of intelligence sources —including informants on the ground— had been established in parts of the historically rebellious, geographically imposing provinces of Kunar and Nuristan. The operators leading the campaign included some of the most highly trained military units at the Obama administration’s disposal, and they were supported by the world’s most powerful electronic surveillance agencies, equipped with technology that allowed for unmatched tracking of wanted individuals.
Despite all these advantages, the military’s own analysis demonstrates that the Haymaker campaign was in many respects a failure. The vast majority of those killed in airstrikes were not the direct targets. Nor did the campaign succeed in significantly degrading al Qaeda’s operations in the region. When contacted by The Intercept with a series of questions regarding the Haymaker missions, the United States Special Operations Command in Afghanistan declined to comment on the grounds that the campaign —though now finished— remains classified.
The secret documents obtained by The Intercept include detailed slides pertaining to Haymaker and other operations in the restive border regions of Afghanistan, including images, names, and affiliations of alleged militants killed or captured as a result of the missions ; examples of the intelligence submitted to trigger lethal operations ; and a “story board” of a completed drone strike. The targets identified in the slides as killed or detained represent a range of militant groups, including alleged members of the Taliban and al Qaeda —but also local forces with no international terrorism ambitions, groups that took up arms against the U.S after American airstrikes brought the war to their doorsteps.
An additional slide included in the materials charts mission statistics from September 2011 to September 2012 for Task Force 3-10, which was responsible for special operations across Afghanistan, breaking down in rare detail the more than 2000 missions conducted by elite U.S. forces in the country over the course of a year.
Together, the materials offer an unprecedented glimpse into the kind of killing that has come to define the war on terror, underscoring the inherent limitations, and human cost, of those operations. With the Obama administration publicly committed to continuing campaigns like Haymaker —special operations missions focused on hunting down specific individuals, not only in Afghanistan but in nations around the world— the documents raise profound questions over the legacy of the longest foreign war in American history and its influence on conflicts to come.
The frequency with which “targeted killing” operations hit unnamed bystanders is among the more striking takeaways from the Haymaker slides. The documents show that during a five-month stretch of the campaign, nearly nine out of 10 people who died in airstrikes were not the Americans’ direct targets. By February 2013, Haymaker airstrikes had resulted in no more than 35 “jackpots”, a term used to signal the neutralization of a specific targeted individual, while more than 200 people were declared EKIA —“enemy killed in action”.
In the complex world of remote killing in remote locations, labeling the dead as “enemies” until proven otherwise is commonplace, said an intelligence community source with experience working on high-value targeting missions in Afghanistan, who provided the documents on the Haymaker campaign. The process often depends on assumptions or best guesses in provinces like Kunar or Nuristan, the source said, particularly if the dead include “military-age males”, or MAMs, in military parlance. “If there is no evidence that proves a person killed in a strike was either not a MAM, or was a MAM but not an unlawful enemy combatant, then there is no question”, he said. “They label them EKIA”. In the case of airstrikes in a campaign like Haymaker, the source added, missiles could be fired from a variety of aircraft. “But nine times out of 10 it’s a drone strike”.
The source is deeply suspicious of those airstrikes —the ones ostensibly based on hard evidence and intended to kill specific individuals— which end up taking numerous lives. Certainty about the death of a direct target often requires more than simply waiting for the smoke to clear. Confirming a chosen target was indeed killed can include days of monitoring signals intelligence and communication with sources on the ground, none of which is perfect 100 percent of the time. Firing a missile at a target in a group of people, the source said, requires “an even greater leap of faith” —a leap that he believes often treats physical proximity as evidence.
The documents include slides focused specifically on Haymaker operations from January 2012 to February 2013, distinguishing between raids, described as “enabled” and “combined” operations, and airstrikes, which are described as “kinetic strikes”. In both cases, raids and airstrikes, the source said the target was always an individual person. “Every mission that’s triggered begins as an objective to find one person for whatever reason”, the source said, adding, “Every jackpot is one person off the list”.
According to the documents, raids performed on the ground during Haymaker were far less lethal than airstrikes and led to the capture of scores of individuals. Research by Larry Lewis, formerly a principal research scientist at the Center for Naval Analyses, supports that conclusion. Lewis spent years studying U.S. operations in Afghanistan, including raids, airstrikes, and jackpots, all with an eye to understanding why civilian casualties happen and how to better prevent them. His contract work for the U.S. military, much of it classified, included a focus on civilian casualties and informed tactical directives issued by the top generals guiding the war. During his years of research, what Lewis uncovered in his examination of U.S. airstrikes, particularly those delivered by machines thought to be the most precise in the Pentagon’s arsenal, was dramatic. He found (1) that drone strikes in Afghanistan were 10 times more likely to kill civilians than conventional aircraft.
“We assume that they’re surgical but they’re not”, Lewis said in an interview. “Certainly in Afghanistan, in the time frame I looked at, the rate of civilian casualties was significantly higher for unmanned vehicles than it was for manned aircraft airstrikes. And that was a lot higher than raids”.
The limited point of view of the drone’s camera, what Lewis describes as the “soda straw effect”, together with the globally dispersed operational network that supports drone strikes, can lead to mistakes, he argues, including the loss of innocent lives. The materials obtained by The Intercept make just one explicit mention of civilian casualties, in the Task Force 3-10 mission statistics from September 2011 through September 2012. The document (2) reveals the U.S. conducted more than 1800 “night ops” at a time when President Hamid Karzai was calling for an end to American involvement in controversial night raids. Of those operations —which resulted in 1239 targets captured or killed and 709 “associates” of targets captured or killed— the military reported “shots fired” in less than 9 percent of its missions, with a total of 14 civilian casualty “events” for the year.
“The 14 civilian casualties is highly suspect”, said the source, who reviewed after-action reports on raids and other operations in Afghanistan. “I know the actual number is much higher”, he added. “But they make the numbers themselves so they can get away with writing off most of the kills as legitimate”.
The Haymaker documents reveal little about whether the deaths reflected in the materials were “legitimate” or not. They do, however, offer an illustrative window into how the killing has been done in the past —and how it may be done in the future.
The request was unambiguous. Dated October 30, 2012, and stamped with the seal of the United States Central Command, the title read, “Request for Kinetic Strike Approval”. The “desired results” listed at the top of the document (3) included just three words : “Kill Qari Munib”.
Munib, whose objective name was “Lethal Burwyn”, was described as a Taliban subcommander operating in the Pech district of Kunar province. He allegedly exercised command and control over a specific portion of the organization, was responsible for numerous attacks on both coalition and Afghan security forces, and communicated with Taliban officials in Pakistan. Specifically, the request reported, Munib had been implicated in recent plots to carry out improvised explosives attacks.
The Americans considered the consequences of taking Munib’s life, including media coverage, possible political fallout, and potential “population blowback”. In all three categories, it was determined that negative repercussions were “unlikely”, and that Munib’s death would “decrease attacks on” coalition and Afghan forces. Going through with the operation, the request asserted, would require a signals intelligence “correlation”, followed by a full motion video lock, visual identification within 24 hours of the strike, and a “low” probability of collateral damage. Two maps were featured in the document intended to seal Munib’s fate, one of which included coordinates of his last known location. In the bottom right hand corner of the document was a bar, numbered one to 10, and fading in color from red to green. It was titled “Confidence Level”. A red triangle sat between the numbers nine and 10.
Less than a week after the briefing was completed, the kill mission was underway. Signals intelligence had been picked up from a compound where Munib was known to sleep, according to a storyboard detailing the operation. Images relayed from the scene revealed the presence of five military-age males in the area. Floating above the site, an MQ-9 Reaper drone, known as “Skyraider”, captured the image of a man dressed in a drab, flowing robe, with a white cap on his head, casting a long shadow in the dirt. According to the storyboard, the image was “correlated” to signals intelligence linked to Munib.
Skyraider loitered above as the man, joined by two others, walked up a ridgeline before heading back into the compound. The figure again emerged from the building. The drone’s camera registered a positive identification. Skyraider “engaged”. A screen grab from the scene shows a cloud of smoke where the individual had been. Task force personnel watched as a group of people collected the target’s remains.
“RESULTS : JP — Pending EKIA, 1 x TOTAL EKIA”, the storyboard reported. An enemy had been killed in action. Confirmation that he was indeed Munib, the jackpot, or JP, was still pending.
On Friday, November 9, 2012, NATO issued one of its standard updates on missions around the country, including two short lines about an operation carried out the day before in eastern Afghanistan. “An Afghan and coalition security force killed Taliban leader Qari Munib during a security operation in Kunar province Thursday”, the statement read (4). “Qari Munib was responsible for directing attacks against Afghan and coalition forces and coordinating the movement of weapons and ammunition for the attacks”.
The announcement appeared five days after the drone strike described in the intelligence community storyboard, which made no mention of Afghan forces involved in the operation. Was Munib killed in a unilateral U.S. drone strike, later obfuscated by NATO ? Or did the drone strike fail to jackpot, resulting in a subsequent joint operation that succeeded in eliminating him ? If so, who was it that Skyraider engaged that day ? Whose body parts did the American analysts watch the first responders collect ?
Those questions remain unanswered. A more fundamental question suggests itself, however. How did the most powerful military in history come to devote its elite forces and advanced technology to the hunt for a man like Qari Munib —a mid-level Taliban figure in a remote corner of the planet, half a world away from the White House and ground zero in Manhattan, more than 11 years after the September 11 attacks ?
When the Americans set out to kill Qari Munib with a drone in 2012, an intelligence document (5) purporting to lay out his bona fides as a target listed local insurgency figures alongside regional actors. In a graphic titled “Link Analysis”, Munib’s name appears under a generic cartoon of an Afghan male, surrounded by six other headshots. Half of them are described as “Salafists”. a conservative faction that has existed in Kunar for decades and, for a period, resisted Taliban presence in the province.
One of the Salafists pictured was Haji Matin, a timber trader from the Korengal Valley. In the early years of the war, one of Matin’s business rivals wrongly fingered him as a militant to the Americans. U.S. forces responded by bombing Matin’s home. While Matin survived, several members of his family were killed (6). The Americans then appropriated one of his lumberyards as an outpost, thus turning one of the most powerful men in the area into a formidable insurgent leader. The transformation of men like Matin and the Salafists, once locally minded powerbrokers, into anti-U.S. fighters, was hardly unique.
“When viewed from absolutely the wrong metric, the Americans were very successful at hunting people”, said Matt Trevithick, a researcher who in 2014 made more than a dozen unembedded trips (7) to some of Kunar’s most remote areas in an effort to understand the province, and American actions there, through the eyes of its residents. The problem, he said, is that savvy, opportunistic strongmen maneuvered to draw U.S. forces into local conflicts, a dynamic that played out again and again throughout the war. “We knew nothing about who we were shooting at —specifically in Kunar”, Trevithick said. He understands the frustration of conventional U.S. forces who were dropped in places like Kunar. “I don’t blame them”, he said. “They’re put in an impossible situation themselves. But what happens is everyone starts looking like the enemy. And that means you start shooting. And that means people actually do become the enemy”.
In September 2010, nine years after the terrorist attacks in New York City, the U.S. military and coalition forces were working their way through a list of 744 people slated for death or capture in Afghanistan. According to the so-called Joint Prioritized Effects List (JPEL), provided by NSA whistleblower Edward Snowden, Kunar, with 44 targets, had the third-highest total in the country. Few of the Kunar targets represented core al Qaeda-linked figures, and many were associated with local groups like the Salafists, whose listed offenses typically included attacks on Western and Afghan government forces in the province. The targets on the 2010 list were prioritized with rankings of one to four, in terms their significance, with one being the most significant. In Kunar, a single target rose to the level of priority one, while more than 80 percent were designated priority three. Seven of the JPEL targets appear in the Haymaker slides, though just three had been linked to al Qaeda. Regardless of their associations, the U.S. ultimately devoted the same resources to picking off locally affiliated militants as it did to the campaign against the group responsible for 9/11.
After nearly a decade of war, thousands of operations, and thousands of deaths, some within the special operations community began to question the quality of the United States’ targets in Afghanistan. “By 2010, guys were going after street thugs”, a former SEAL Team 6 officer told (8) the New York Times recently. “The most highly trained force in the world, chasing after street thugs”. Concerns that the U.S. was devoting tremendous resources to kill off a never-ending stream of nobodies did little to halt the momentum.
The secret documents obtained by The Intercept, which include a slide (9) on “Manhunting Basics”, reflect the combination of U.S. military personnel and spies who have hunted targets along Afghanistan’s border with Pakistan for years. According to one of the slides (10), the Haymaker “functional teams” included the CIA, the Defense Intelligence Agency, the NSA, and the National Geospatial-Intelligence Agency. The Air Force’s uniquely designed 11th Intelligence Squadron also played a role. The Florida-based squadron was reactivated (11) in August 2006 for the express purpose of supporting “find, fix, finish” operations to capture or kill targets through analysis of aerial intelligence.
The drone operations that supported campaigns like Haymaker also included personnel stationed at Camp Alpha, a secure facility at Bagram populated by teams from the military’s Joint Special Operations Command (JSOC), as well as contractors manning stations at U.S. bases like Fort Gordon, a lesser-known though crucial node in the war on terror that supports tactical NSA operations abroad from Augusta, Georgia. The hunting and killing operations relied on advanced technology to zero in on targets, including the cellphone geolocation system known as GILGAMESH. As The Intercept reported (12) in 2014, the GILGAMESH system employs a simulated cellphone tower to identify and locate targeted SIM cards.
While signals intelligence and electronic surveillance clearly contributed to Haymaker’s kinetic operations —like the drone strike that targeted Qari Munib— there was evidently more to the missions than advanced technology. Unlike some other arenas in which the war on terror has touched down —Yemen or Somalia, for example— the documents point to the robust presence of U.S. intelligence agencies and human sources on the ground in northeastern Afghanistan. In Nuristan’s Waygal district, the Defense Intelligence Agency and the CIA had “myriad complementary sources”, a slide (13) laying out “targeting criteria” in the district noted, “some of whom may be able to trigger our operations”. A third of the “active targets” in Waygal had “good selectors” —phones the Americans could target in the run-up to a raid or airstrike— and the NSA had “taken on [signals intelligence] development in Waygal, greatly enhancing our understanding of the [signals intelligence] environment”.
According to the slide, Waygal, described as a “historic” al Qaeda sanctuary, included more than a half-dozen NAIs, “named areas of interest”, the identification of which was attributed to similarly “historic” levels of human and signals intelligence cultivation, as well as surveillance provided by drones scanning the district. There were “over a dozen active targets” in three villages, the slide said, adding that most of the targets were already on a targeting list, or “easily could be”. “The targets there are not only senior-level Taliban facilitators and hosts, but Arabs themselves”, the slide noted, underscoring the presence of suspected foreign fighters in the district. “Elimination of these targets”, it continued, “will provide demonstrable measures of success”.
The documents indicate that U.S. forces launched just one airstrike as part of the Haymaker campaign in the early months of 2012, killing two people. In May 2012, however, the tempo of operations picked up dramatically, an increase that coincided with a strategic shift in Afghanistan emanating from the White House. As the military’s focus shifted to hunting down specific targets from 2011 to 2012, drone strikes in Afghanistan increased (14) by 72 percent.
Over the course of five months, stretching through the summer of 2012, Haymaker operations included 27 raids and 27 airstrikes. The raids resulted in the capture of 61 people, 13 of them jackpots, the actual targets of the missions. A total of two people were reportedly killed in these ground operations. In the airstrikes, meanwhile, a total of 155 people were killed and labeled as enemies killed in action, according to a table (15) presented in the documents. Just 19 were jackpots. The table does not say whether the jackpots are reflected in the EKIA total. It does, however, appear to present a success rate : the number of jackpots divided by the number of missions. In the case of raids, a figure of 48 percent is presented ; for airstrikes it’s 70 percent.
The scores of unnamed people killed in the hunt for jackpots, and the intelligence opportunities lost by failing to capture targets alive, do not appear to factor into the calculation. The apparent success rate, in other words, depends solely on killing jackpots, and ignores the strategic —and human— consequences of killing large numbers of bystanders.
While the source conceded there could be scenarios in which women and children killed in an airstrike are labeled as EKIA, in the case of the Haymaker strikes he believed it was “more likely” that the dead included “groups of men or teenaged boys” killed because “the intel says the guy JSOC is going after may be in that group of men”. In the event that a target is identified in such a group, he said, “They’ll go through with the strike”.
The materials also include a chart revealing that airstrikes killed 219 people over a 14-month period in 2012 and 2013, resulting in at least 35 jackpots. The document includes (16) thumbnail images of individuals, representing a range of groups, who were captured or killed during Haymaker —in total 30 men, 24 of them stamped EKIA, five detained, and one wounded in action. The deaths of just over half the individuals were noted in NATO’s press releases or media reports.
The Haymaker files also point to the psychological impact of living under the constant threat of death from above —an effect human rights workers have documented (17) among civilians living in areas populated by militants. A quote attributed to a Taliban detainee identified as “Ahmad”, aka “Objective Brandywine”, features prominently on three of the documents (18). “Hands down, the scariest/most intimidating message for the Taliban, at any level, from fighter to Taliban senior leadership, is anything to do with drones or aerial bombings”, Ahmad purportedly said. “The Taliban has no way to defend against them and they are certain to end in absolute destruction of whatever their target is”.
Still, the documents’ assessment of Haymaker’s effectiveness was frank. A slide detailing the campaign’s “effects” from January 2012 through February 2013 included an assessment of “Objectives & Measures of Effectiveness”. The results were not good. Disruptions in al Qaeda’s view of northeastern Afghanistan as a safe haven and the loss of “key” al Qaeda members and enablers in the region were deemed “marginal”. Meanwhile, a comparison (19) of Haymaker 1.0 (August 2011) with Haymaker 2.0 (February 2013) noted that al Qaeda faced “little to no local opposition” and enjoyed “relatively free movement” to and from Pakistan. Kinetic strikes, the slide reported, “successfully killed one [al Qaeda] target per year”, allowing the organization to “easily” reconstitute.
By 2013, Haymaker was amassing a significant body count but making little headway against al Qaeda forces in the region. According to the “Success Criteria” slide, “sporadic reporting of concern over [the] viability” of northeastern Afghanistan as a safe haven had been “overshadowed” by the group’s senior leadership discussing the establishment of a “post-2014 sanctuary”. Individuals continued to return to Pakistan to support operations in and outside of Afghanistan, the slide asserted. While “nascent developments in some valleys” indicated that locals were “tiring” of al Qaeda’s efforts to “root out spies as a perceived method to stopping strikes”, the strikes and raids themselves had “succeeded in killing/capturing few [al Qaeda] targets”. As slides detailing its effectiveness noted, Haymaker’s impact on al Qaeda and Taliban enablers in Kunar and Nuristan was “considered temporary without a long-term, persistent campaign”.
On February 18, 2013, while Haymaker was still underway, Afghan President Hamid Karzai issued a decree (20) : Afghan military forces were barred from calling in U.S. airstrikes for support on missions. The order followed an operation in Kunar in which NATO and Afghan forces were blamed for the deaths of 10 civilians —including one man, four women, and five children. In 2012, the U.N. mission in Afghanistan had documented a number of other incidents involving civilian deaths resulting from U.S. operations, including a raid that left seven civilians dead, an “aerial attack” that killed seven children and one adult, and a drone strike targeting “two insurgents” that killed a teenage girl.
The most recent date included in the Haymaker materials is February 28, 2013. Whether the date marked the end of the campaign is unclear. What does seem clear, however, is that Haymaker coincided with an increase in drone strikes and civilian casualties across Afghanistan. By the end of 2013, the U.N. reported the number of civilian casualties from drone strikes in Afghanistan had tripled from 2012, with “almost one-third of the civilian deaths from aerial operations” reported in Kunar, the heart of the Haymaker campaign. Records of condolence payments disbursed by the U.S. military, obtained (21) by The Intercept, show more than $118000 distributed in 45 disbursements to Kunar in fiscal years 2011 through 2013. In addition to numerous injuries, the payments also cover the deaths of 27 people, including at least four children. The records do not indicate whether the incidents were linked to the Haymaker campaign or whether they were the result of mistaken ground raids or airstrikes.
Until recently, the ongoing conflict in Afghanistan had largely receded from public conversations in the U.S. This month, an American airstrike on a hospital run by the international organization Médecins Sans Frontières, offered a forceful reminder that the war, despite the Obama administration’s declaration in 2014, is far from over. Unleashed in the early morning hours of October 3, in the province of Kunduz, the U.S. attack killed at least a dozen members of the humanitarian group’s medical staff and 10 patients, including three children. A nurse on the scene recalled seeing six victims in the intensive care unit ablaze in their beds. “There are no words for how terrible it was”, the nurse said (22). MSF denounced the strike as a war crime and demanded an independent investigation.
The Kunduz attack underscored (23) an ugly reality : After nearly a decade and a half of war, more than 2300 American lives lost, and an estimated 26000 Afghan civilians killed (24), the nature of combat in Afghanistan is entering a new, potentially bloodier, phase. In August, the United Nations reported (25) that civilian casualties in Afghanistan “are projected to equal or exceed the record high numbers documented last year”. While most civilian casualties in the first half of 2015 were attributed to “anti-government” forces, 27 deaths and 22 injuries were attributed to airstrikes “by international military forces”, a 23 percent increase over last year, most of them, unlike the air raid in Kunduz, carried out by drones.
Despite the rise in civilian casualties and the well-documented failure of drone strikes to achieve the military’s broader objectives, there is every indication that unmanned airstrikes will play an increasing role in U.S. military engagement in Afghanistan, as they have in war zones across the world. Less than two weeks after the U.N. issued its report, Foreign Policy revealed (26) that JSOC has drastically reduced the number of night raids it conducts in Afghanistan, while dramatically increasing its reliance on airstrikes, and is currently “on pace to double the rate at which it kills "high-value individuals" using kinetic strikes, compared to how many it was killing that way five years ago”.
Afghanistan’s northeastern border with Pakistan remains an active area of focus for the remaining U.S. special operations forces in the country. The Pech Valley, once a hotspot during the Haymaker campaign, continues to host a constellation of armed groups. Al Qaeda, the organization used to justify both the invasion of Afghanistan and the Haymaker campaign, reportedly enjoys a more pronounced presence in the valley than ever. “The al Qaeda presence there now”, according to a report (27) by the United States Institute for Peace, “is larger than when U.S. counterterrorism forces arrived in 2002”.
With JSOC and the CIA running a new drone war (28) in Iraq and Syria, much of Haymaker’s strategic legacy lives on. Such campaigns, with their tenuous strategic impacts and significant death tolls, should serve as a reminder of the dangers fallible lethal systems pose, the intelligence community source said. “This isn’t to say that the drone program is a complete wash and it’s never once succeeded in carrying out its stated purpose”, he pointed out. “It certainly has”. But even the operations military commanders would point to as successes, he argued, can have unseen impacts, particularly in the remote communities where U.S. missiles so often rain down. “I would like to think that what we were doing was in some way trying to help Afghans”, the source explained, but the notion “that what we were part of was actually defending the homeland or in any way to the benefit of the American public” evaporated long ago. “There’s no illusion of that that exists in Afghanistan”, he said. “It hasn’t existed for many years”.