It was January 2002 in Afghanistan. Snow covered the mountains that surrounded Bagram air base, where, in a frigid electrical closet, FBI special agent Russ Fincher and New York Police detective Marty Mahon were interrogating Ali Abdul Aziz al-Fakhri, a Libyan known by his nom de guerre, Ibn al-Shaykh al-Libi. One of the highest-level Al Qaeda figures captured by that point in the war, Libi had run the group’s Khalden training camp in eastern Afghanistan. Fincher and Mahon built a relationship of trust with Libi, who talked freely. In particular, he divulged what a military source who was in Bagram at the time described as Al Qaeda’s “multi-phased” plans to regroup after being forced from its safe haven in Afghanistan. The first phase was to flee across the border to Pakistan’s tribal areas, but to be prepared for further movement. Assuming they would not be safe for long in the tribal areas, Al Qaeda leaders’ ultimate goal was to reconstitute their force in the next best potential sanctuaries: Yemen and Somalia.1
History and geography argued in favor of believing Libi. East Africa and Yemen had been the sites of Al Qaeda’s most sensational attacks prior to September 11, 2001. On August 7, 1998, the group staged almost simultaneous truck bomb attacks on the U.S. embassies in Nairobi, Kenya, and Dar es Salaam, Tanzania, killing more than 200 people, the vast majority of them locals. On October 12, 2000, Al Qaeda conducted a suicide boat attack on the Cole, a U.S. Navy destroyer, killing seventeen sailors and blowing a hole in the ship while it was at harbor in the Yemeni port of Aden. The United States knew, therefore, that Al Qaeda already had roots in the region. It also knew that Yemen was Osama bin Laden’s ancestral homeland.
Sure enough, in 2002 U.S. intelligence noticed small numbers of second-tier Al Qaeda figures moving back and forth between the Afghanistan-Pakistan theater and the Horn of Africa and Yemen, traveling by boat from Oman, past Yemen, and across the Bab-el-Mandeb Strait, which separates the Red Sea from the Gulf of Aden, before following the coast of Djibouti down to Somalia. When Al Qaeda also began transferring money into the region—a key indicator that operational planning might be occurring—U.S. leaders became alarmed.
But with a war under way in Afghanistan and another planned for Iraq, the Bush administration decided to wage its campaign against Al Qaeda’s East African and Yemeni branches largely in the shadows, using the two weapons upon which it would increasingly rely in the years ahead: the CIA and JSOC.2 For the first four years of the campaign in the Horn, the CIA would take the lead. The Agency gave the campaign a name that recalled JSOC’s past experience in Somalia: Operation Black Hawk.
* * *
Starting in early 2002, small teams of U.S. operatives began to conduct missions into Somalia. The first trip was by car, from Addis Ababa, the Ethiopian capital, to Hargeisa, Somalia’s second largest city and the capital of the autonomous Somaliland region in the country’s northeast. Shortly thereafter the U.S. operatives began flying to Baidoa, a city in southwest Somalia. On these first trips the operatives would start their missions unarmed, before renting AK-series assault rifles once they were on the ground in Somalia. In 2003 the operations center for the clandestine campaign in Somalia switched from Addis Ababa to the Kenyan capital of Nairobi. The teams would drive to Nairobi’s Wilson Airport and climb aboard a chartered Bluebird Aviation turboprop that would fly them to the K50 Airport, about fifty kilometers southwest of Mogadishu. Although two Bluebird planes would crash into each other a couple of years later, it’s still the case that for the Americans aboard, the flights were probably the least dangerous part of missions that demanded professionalism, courage, and coolness under pressure.
In those early days, the teams combined at least two CIA case officers, two Army of Northern Virginia operatives, and an interpreter. Once the planes landed, the teams would travel to and through Mogadishu in small convoys escorted by fighters loyal to one warlord or another. Different patches of the anarchic city were controlled by different warlords, requiring much coordination to ensure safe passage as the convoys crossed the boundaries between the warlords’ territories.
Those warlords were the key to Operation Black Hawk. The CIA was paying them to kill or capture the twenty or so most senior members of Al Qaeda’s East Africa cell. If the warlords captured one of these targets, they were to turn him over to the Agency, which would send—or “render”—him to a U.S. ally or one of the CIA’s secret prisons. The man in charge was John Bennett, the Agency’s highly regarded Nairobi station chief, a former Marine infantry officer who in 2010 would become the head of the National Clandestine Service—the CIA’s top spy. Bennett did not travel to Mogadishu, but his leadership was critical to the effort. “The relationship with the warlords was built through … Bennett,” said an intelligence source familiar with the missions. “It was through his sheer willpower and personality. He could do it and nobody else could.”
Bennett also worked well with, and was respected by, JSOC. On those first missions, the military personnel were Army of Northern Virginia operatives, whose primary role was to provide security as the CIA gathered and validated human intelligence. But the operatives came into their own as they began to install gear around Mogadishu to monitor the city’s cell phone traffic, which NSA satellites could not capture. “The problem was you cannot do intercept through a lot of national asset capability,” said an officer familiar with the unit’s operations in the Horn. “You have to be there on the ground, for a lot of this stuff is very tactical in nature, it’s very temporal. You have to be Johnny on the spot.” Just as in Syria, these devices, which were sometimes placed in warlords’ homes, had to be serviced and moved regularly as new cell phone towers sprang up and old ones were repositioned. The CIA had its own signals intelligence capability, which it jealously guarded, but it was keen to augment its tiny fleet of signals intelligence aircraft in the region with a modular signals intelligence package that could be rolled on and off a rented plane. The Agency wanted the package to collect cell phone traffic from an airborne position up to four or five miles from the target area. However, significant technical difficulties were involved in making such a system work without integrating it into the plane’s airframe. Developing the capability took several years, and the help of Orange (as the Army of Northern Virginia was now known), after which the modular capability was made available to both organizations.
The CIA used a carrot-and-stick approach to working with the warlords, handing them suitcases full of hundreds of thousands of dollars, but with the implicit threat of U.S. air strikes if they betrayed the United States. “They were risky missions,” said the intelligence source. “You could never actually trust the warlords—they’re subject to the highest bidder.” But the Americans were bluffing about the air strikes. No U.S. aircraft were nearby, not even drones to monitor the missions. The escape and evasion plan in case a team got into trouble was “Get to the coast and hope for the best,” said a JSOC staffer. At the time, the staffer noted, the Navy had no ships in the Indian Ocean, and the closest “fast-mover” jets that might provide close air support were five hours away, in Qatar.3
After Bennett left in August 2003, the CIA station’s focus began to shift and its appetite for risk waned, according to a JSOC source stationed in Nairobi. For two years, the CIA operatives who flew into Mogadishu never left the plane, holding their meetings with the warlords on the Bluebird. During one such meeting, the warlord pointed out of the window to four white female reporters going about their business. His point was clear: Those unarmed Western women aren’t scared to work in Somalia—why are you? The three-man JSOC teams that accompanied such flights but were also confined (on the ambassador’s orders) to the aircraft were now filled by a couple of Team 6 operators in addition to an Orange signals intelligence soldier. Ironically, the Agency treated the operators as “hired guns,” but banned them from bringing rifles on the missions, said a SEAL. The Orange operative would either position his phone-monitoring device at the airport near Mogadishu or would have an Agency source emplace it.4
While the CIA was in overall charge of the Black Hawk missions, the JSOC personnel reported to a Team 6 or Delta officer in Nairobi who oversaw a small interagency team established in mid-2003 under JSOC auspices, staffed with intelligence and law enforcement personnel and working out of two small rooms in the embassy. He in turn worked for the Orange commander, who between 2003 and 2005 was Konrad Trautman. (Just as McChrystal had given Iraq to Delta, he had placed Orange in charge of the Horn and Yemen.) JSOC doubled its tiny resources in the Horn between 2003 and 2005, focusing more tightly on intelligence collection and target development. These efforts meant “we gained a lot of understanding of what was going on,” a senior intelligence official said.5 McChrystal began conducting Horn of Africa–specific video-teleconferences that connected operators, ambassadors, and CIA station chiefs with officials in Washington. He also placed a small team in Addis Ababa as part of this effort to thicken JSOC’s network in the Horn. JSOC was seeking to target Al Qaeda and its associates in the Islamic Courts Union, an Eritrean-financed Islamist force that had gained control of much of southern Somalia. Conversely, by the middle of the decade, the Nairobi CIA station seemed more interested in collecting intelligence for its own sake than on hunting Al Qaeda. “The Black Hawk team was not interested in CT [counterterrorism], they were interested really in foreign intelligence collection,” said a Team 6 operator, adding that the warlords got the better part of the deal. “They were paying these warlords vast sums of money for nothing,” he said. “From my perspective it was a complete failure. I’m sure from the Agency’s perspective it was not.”
(A JSOC staffer took issue with the operator’s account, and said Team 6 deserved some blame itself. Orange “had safe houses all over Somalia by the time Blue came in,” he said. “All that went [away].” A U.S. intelligence officer, meanwhile, disputed the notion that the CIA’s Nairobi station became less interested in counterterrorism after Bennett’s departure. “Every day [the new chief of station] was hammering, ‘What have we done on CT?’” he said.)
The CIA had also decided that there weren’t enough targets being developed in Somalia to keep its Predator fleet in nearby Djibouti, a tiny Muslim country that had agreed to host U.S. forces. The drones went away.6 But the U.S. military was building up its forces in Djibouti, creating Combined Joint Task Force–Horn of Africa in November 2002. That task force ostensibly focused on civil affairs missions and on strengthening the capacity of allies in the region. It also gave the small but growing JSOC presence in Djibouti a larger organization in which to hide.7
In late summer 2005, the Black Hawk teams got permission to occasionally get off the planes and go into Mogadishu again. The CIA tried to extend the rule that banned rifles on the planes to the missions in which operators—from Team 6, Delta, and Orange—disembarked and went into Mogadishu, sometimes staying overnight at the residence of Bashir Raghe Shiiraar, a secular warlord. The Agency argued that the warlord militias would provide enough protection. An incident at a Somali airfield soon exposed the emptiness of that promise. A Black Hawk team that included a couple of Team 6 operators was sitting on the plane for a warlord meeting when a rocket-propelled grenade flew across the runway. The warlord’s forces who were supposed to be protecting the Americans disappeared. The SEALs, who had been carrying their rifles broken down in their packs, quickly put the weapons together, got off the plane, and seized the highest ground they could, which was a berm around the airfield. Whoever had fired the RPG disappeared. The CIA officers said nothing about the weapons.8
The friction between the Agency and JSOC in the Horn extended to the highest levels. In 2005 McChrystal visited the embassy in Nairobi. The station chief let it be known ahead of the visit that McChrystal “had better be prepared to get down on bended knee.” When the meeting took place, the station chief, a short man with a “huge ego,” according to a source who worked close to him, condescended to the JSOC commander. McChrystal, who as Ranger Regiment commander had overhauled the regiment’s hand-to-hand combat program, sat and listened. “When the guy stopped talking, McChrystal finally says, ‘Hey look, if you ever talk to me that way again, I’m going to come around this desk and beat the shit out of you,’” said a source who was in the room. “And that changed the whole tone between the two of them. It became that old cartoon Spike and Chester, Spike’s the bulldog and Chester’s always jumping over him saying, ‘Can I be your buddy, Spike, can I be your buddy?’ All of a sudden [the station chief] now wants to be McChrystal’s buddy. It was disgusting actually but I’m thinking to myself, ‘That’s all it would have taken—just threaten the guy physically.’ Classic move.”
By late 2005, the Orange operatives’ courage and professionalism had earned the respect of their peers in the Horn. “Orange did a great job,” said a Team 6 operator. At any one time, there were between two and six operatives—i.e., those personnel who went into Somalia with the signals intelligence gear—in Nairobi. They were there on typical JSOC three-to-four-month rotations, using fake last names with thin official cover while there. Orange had selected a few personnel for the mission whose ethnicity would not draw attention, meaning a greater range of covers was available to them. “A lot of them you wouldn’t recognize as Westerners,” said a special mission unit member. For instance, one operative was a black American who, in civilian clothes, could pass as an African and spoke fluent Swahili, he said.
Of course, these covers came with their own risks. During the 2005 to 2006 winter, an operative who resembled a member of Kenya’s ethnic Indian merchant class was shot in the abdomen late at night at a gas station as he filled the tank of his SUV in an affluent part of Nairobi near the U.S. embassy. The operative was rushed to a nearby hospital and survived. “Of course, this sent everyone spinning to [figure out], have we been compromised? Are people following us?” said the special mission unit member. However, in part because the robbers stole the operative’s wallet, the consensus was that he was the unfortunate victim of a crime of opportunity.9
With every trip into Mogadishu’s urban jungle, the Orange operatives strengthened the United States’ ability to keep tabs on its enemies in the Horn. But events beyond JSOC’s control would soon undo much of that work, while ultimately offering the command new opportunities.
* * *
As with the Horn, JSOC’s post-September 11 presence in Yemen started small. In summer 2002, three Delta operators under official cover, including a male-female team, arrived in the country on a classic advance force operation. Their job was to begin to gain an understanding of the political-military environment and how transnational factors were affecting it.10 Two Army of Northern Virginia operatives had actually preceded them, arriving in February 2002. By the end of the year, the unit had been placed under JSOC and become Task Force Orange, but the plan to ramp up the military’s clandestine force in Yemen had already suffered a couple of setbacks.
First, Yemeni officials seized some of the Army of Northern Virginia’s critical signals intelligence technology when it arrived at Sana’a airport, hidden in a larger shipment of gear for U.S. theater (i.e., non-JSOC) special operations forces that were training Yemeni troops. The kit was essential for the unit’s mission, but the U.S. embassy refused to pay the import duty that the Yemenis demanded, in the belief that it was a shakedown. Complicating matters was the fact that neither the Yemeni officials nor U.S. ambassador Edmund Hull understood what the gear was for. “He wasn’t cleared for that,” said a military source with Yemen experience. (In his memoir, Hull referred to the Yemenis refusing entry to pallets that the U.S. government considered to enjoy the same protection as diplomatic pouches, and listed “radios, weapons, and blood” as among the “equipment” items the pallets contained.) Following the loss of the gear, JSOC personnel were involved in a car crash in early summer 2002 that killed a Yemeni. They had to leave the country. It took several months, Hull’s involvement, and the payment of what the military source termed “blood money” before JSOC was allowed back in.
Meanwhile, Hull was resisting what he described as “a great deal of pressure” from Washington for unilateral U.S. combat missions in Yemen. The ambassador opposed air strikes on the grounds that they would “inevitably” result in high numbers of civilian casualties. He acknowledged that special operations forces might “theoretically” provide “a more surgical option,” but added that “planning always entailed options for massive support for the special [operations] forces should they become trapped.” As with its proposal for attacking Ansar al-Islam’s camp in Iraqi Kurdistan earlier that year, JSOC’s insistence on a massive JRX-style operation became an insurmountable stumbling block. By the fall, JSOC’s numbers began to climb again, but in September it only had six or seven personnel in Yemen. The U.S. counterterrorism mission in Sana’a at the time was clouded by a multitude of actors and differing chains of command.11 There was a team from Central Command’s special operations headquarters, which commanded the theater special operations forces; the small JSOC element; the CIA station; and a national intelligence support team that included a cryptologic support group from the National Security Agency.12 The JSOC and intelligence personnel were trying to locate Al Qaeda targets, particularly those associated with the Cole attack. But there was competition between the CIA and the military over who could get permission to strike first, if one of those targets hove into view. “The Agency was a little faster,” said the military source with Yemen experience. “Probably a lot faster.”
This became clear on November 3, 2002, when the NSA detected a call made on a phone associated with Qaed Salim Sinan al-Harethi,13 a Yemeni Al Qaeda member who was suspected of helping to plan the Cole attack. The United States had been trying to locate Harethi all year using cell phone tracking technology. They had located him at least once before that summer, but while the U.S. and Yemeni governments were planning a strike, the United States monitored a call to Harethi’s phone from the Yemeni defense ministry, warning him. Not surprisingly, he disappeared and stopped using that phone. “He went chilly for about three months,” said the military source. Now, for whatever reason, the experienced jihadist had chosen to use it again. This time the Agency was taking no chances. A Predator flew from Djibouti and destroyed the car, which was being driven through the desert in Marib province, about 120 kilometers east of Sana’a, killing Harethi and five fellow passengers. It was the first known lethal drone strike outside Afghanistan.
Orange kept a small team in Yemen, operating out of the embassy, but the next few years were relatively quiet from a counterterrorism point of view. That began to change on February 4, 2006, when twenty-three Al Qaeda members tunneled their way to freedom from the Political Security Headquarters jail in Sana’a.
* * *
Meanwhile, back in the Horn, Orange’s signals intelligence missions were paying off. “It definitely led to us being able to have much more precise information about what was going on,” said the senior intelligence official. “Those operations gave us pretty good insight into what Al Qaeda was doing in East Africa. They saw it as another safe haven, they saw the opportunity to establish training camps, and they did. And it allowed us to start to plan [counterterrorism] operations against a couple of the key targets.”14
The gear that the operatives used was not much to look at—“essentially it’s a box and an antenna”—but emplacing it required significant expertise, said the special mission unit member. “You just can’t put it anywhere to get the collection you want. So you really have to go in and survey both the physical geography and the electron environment.” Figuring out the best places “to collect against your signals of interest” was laborious and time-consuming work. “Electronic communications engineers, they love that sort of problem,” he said. “But if you’re in a place like Mogadishu, it’s a dangerous place to be out, figuring that stuff out.” However, the need to safeguard the technology meant some otherwise optimal locations were ruled out. “You really want it in a controlled place,” the special mission unit member said. “You don’t want to lose that stuff.” On some occasions, the operatives settled for “the least bad place to put them,” he said. Of course, once the machines were up and running, someone had to translate the intercepted phone calls. Orange used a combination of its own personnel and contractors—“just phenomenal linguists” who not only understood Somali dialects but “had been doing it so long they could immediately recognize targeted or key individual voice characteristics,” the special mission unit source said.
Despite JSOC’s increasing ability to track targets in Somalia, the command mounted no successful air strikes or raids into the country during the first half of the decade. During that period, warlords paid by the CIA helped render “seven or eight” Al Qaeda figures out of Somalia, said an intelligence source with long experience in the Horn. Reluctant to put these detainees on trial in the United States, for fear of divulging the intelligence “sources and methods” that led the CIA to them, the Agency transferred at least some of them to its “Salt Pit” secret prison in Afghanistan.15
While Orange’s technologies helped locate Islamists targeted for rendition, the CIA’s warlord allies were letting it down on the wider battlefield. The Islamic Courts Union steadily gained ground during the first half of 2006, finally taking control of Mogadishu the first week of June, when it ran the secular warlords out of town.16
The warlords’ defeat was a disaster for U.S. policy, for the CIA’s strategy, for JSOC’s operational ambitions, and for Orange’s signals intelligence program. Not only did Orange—and therefore JSOC and the National Security Agency—lose access to the locations where it had been setting up its devices, it also lost a couple of the devices themselves when the Islamic Courts Union fighters overran the positions in which they’d been positioned. The devices, of which Orange had fewer than ten in its inventory, were not disguised. “This is clearly Western intelligence agency equipment, which was the biggest concern,” said the special mission unit member. “So losing those was not good.” The loss of the kit not only meant “a loss of collection,” but also that the United States’ enemies in Somalia, realizing that they were being listened to, might start engaging in deception when talking on their cell phones. “So you’ve got all sorts of problems created by that,” he said. It took about a year for JSOC to regenerate its signals intelligence capability in Somalia.17
Part of JSOC’s problem was that, just as Dave Schroer had predicted to the Defense Department leadership in early 2003, the demands of the Iraq War had forced JSOC to neglect the Horn of Africa. Now, three years after the invasion of Iraq, the growing Al Qaeda presence in the Horn and Yemen so alarmed senior military and intelligence leaders that they were prepared to take risks in Iraq and Afghanistan in order to deploy more ISR assets to the Horn.18 In late 2006, however, the JSOC presence in Kenya consisted of no more than a dozen military personnel: two or three Team 6 operators; a technical surveillance equipment support service member, also from Team 6; one or two Orange signals intelligence squadron personnel; a Joint Communications Unit radio expert; and an officer to command the team, which operated out of the Nairobi embassy. Only the SEALs and the Orange personnel would go into Somalia.19
* * *
The Team 6 operators were in trouble and they knew it. There were just a couple of them, plus a 24th Special Tactics Squadron combat controller, embedded in a larger unit of the security services of northeastern Somalia’s autonomous Puntland region, and they had bitten off more than they could chew. It was June 1, 2007. They had been on the hunt for a multinational group of Al Qaeda fighters who had arrived a few days earlier in the Somali town of Bargal, on the very tip of the Horn of Africa, having traveled up the coast from southern Somalia. But the hunters had become the hunted. Outnumbered and outgunned, the operators turned to the combat controller as a last resort.20
The operations that led up to the Battle of Bargal began in late December 2006, when Ethiopia invaded Somalia in response to the growing strength of the Islamic Courts Union. (Ethiopia’s traditional enemy, Eritrea, had been funding and supplying the Islamists.) In a December 6 cable, U.S. Ambassador to Ethiopia Donald Yamamoto accurately predicted that Ethiopia would invade later that month and that the incursion might “prove more difficult for Ethiopia than many now imagine.” But oddly, although JSOC had been building up a small force at Dire Dawa in eastern Ethiopia throughout late 2006, presumably in preparation for combined operations with the Ethiopian forces in Somalia, the invasion took JSOC by surprise. “We should have been leaning forward to capitalize on this, and we did nothing,” said a senior military official. JSOC was forced to scramble. It took until late March or April to deploy about a dozen operators to link up with the Ethiopians in Somalia. Most were from Team 6’s Gold Squadron, along with a few Delta and 24th Special Tactics Squadron operators. Split into two- and three-man teams and inserted into Ethiopian infantry units, their mission was to advise and assist the Ethiopian forces, who were ousting the Islamists from Mogadishu and driving them south toward the Kenyan border. Even that tiny deployment, which required Rumsfeld’s approval, created heartache among Washington policymakers who were fearful of a repeat of 1993’s costly Mogadishu battle. “It was very uncomfortable,” the intelligence official said, adding that if McChrystal had had his way, JSOC “would have gone with a much bigger capability and been much more aggressive.”21
The command also stationed two AC-130 gunships at Dire Dawa, to which the Ethiopian government consented on the condition that their missions and presence on Ethiopian soil remain secret. At least one gunship was soon in action, striking a column of suspected Islamic Courts Union and East African Al Qaeda fighters on January 7 near the port of Ras Kamboni in southern Somalia. The main target of the attack, Aden Hashi Ayro, leader of the Islamic Courts Union’s al-Shabaab militia, was wounded but survived and escaped. Another series of air strikes in the same area followed two days later. But a third mission near the Kenyan border a little more than two weeks later proved to be the gunship deployment’s undoing. The early morning strike targeted Ahmed Madobe, a deputy of Islamic Courts Union head Hassan Turki, but only succeeded in wounding him and killing eight of his companions. A few hours after the strike, a helicopter landed carrying Ethiopians and Americans who seized him and took him to a facility near the Somali city of Kismayo, where they interrogated him and treated his wounds.
The Washington Post reported on the raid on January 24, infuriating the Ethiopian prime minister, Meles Zenawi, who had stressed the need for operational security and urged the United States to keep a “light footprint” in the area. Meles asked that gunships leave Ethiopia and for the United States not to engage in military strikes in Somalia, but to pass targeting information to his forces instead. In a January 25 cable, Yamamoto said he agreed with Zenawi that the gunships should depart. “Heavy press interest has made it difficult to secure and protect the AC-130 operations,” he wrote. The gunships left shortly thereafter.22
Planned as a quick in-and-out operation to dislodge the Islamists, the Ethiopian invasion soon bogged down into an occupation. The JSOC operators spent only a few weeks with the Ethiopians before pulling out, their most serious casualty a SEAL who came down with malaria. By then, JSOC had opened an “outstation”—a small base or safe house from which a handful of operators and support personnel worked with local forces—in Bosaso, a port on Puntland’s northern coast. “That was a relationship that started with the Black Hawk teams that we pretty much took over,” said a Team 6 source. It was a team from Bosaso that found itself pinned down in Bargal on June 1. Out of other options, the 24th STS combat controller called the one source of U.S. firepower nearby: the Chafee, a Navy destroyer off the coast. The Chafee fired more than a dozen rounds from its five-inch gun. That naval gunfire—rare in the twenty-first century—enabled the U.S. and Ethiopian troops to break contact and get away.23
JSOC’s missions in Somalia were taking on a lethal aspect. The aim increasingly became to kill targets, rather than to capture them. But that shift required a much bigger support presence in Nairobi to enable JSOC’s small teams in Somalia. Accordingly, the task force presence in the embassy quickly grew to about seventy people as intelligence analysts and other support personnel arrived. But as 2007 wore on, the relationship between JSOC and the Ethiopian military began to fray. This was in part because the Ethiopians did not want to be seen as U.S. proxies, but also because the priorities of the United States and Ethiopia overlapped, but were not the same. Ethiopia’s primary goals were to oppose the threat posed by the Islamic Courts Union and to prevent its bitter enemy Eritrea from using the group as a proxy to attack Ethiopia. The United States was focused on killing a handful of people at the top of the East Africa Al Qaeda cell, and had little interest in killing large numbers of Islamist foot soldiers. “If we wanted to kill a couple of thousand guys, we could have done that pretty much any time,” said the senior intelligence official.24
The Ethiopian invasion essentially reinstalled Somalia’s Transitional Federal Government, but the TFG controlled very little territory and was wholly reliant on its foreign backers. Meanwhile, from JSOC’s perspective, the situation in Somalia was worsening. Put another way, the more force JSOC applied to the problem in Somalia, the more work the command found for itself. U.S. intelligence concluded that up to 300 Islamist fighters arrived in Somalia in summer 2007.25 JSOC believed that a similar number of militants were training in just two camps near Ras Kamboni.26 Operators also accompanied Kenyan forces to the border in order to help the Kenyans intercept senior Islamists trying to slip into their country.27
Although the decision to move assets from Iraq and Afghanistan to the Horn meant there were now Predators and manned ISR aircraft flying over Somalia from Djibouti, it was nowhere near the sort of coverage to which JSOC was accustomed. In contrast to JSOC’s “Unblinking Eye” in Iraq, “in Somalia, it was a blink all the time,” the senior intelligence official said. There would be days on end when task force commanders in the Horn had “no overhead collection capability,” the official added.28
The lack of airborne signals and imagery intelligence collection might not have been so painful for JSOC had the Defense Intelligence Agency not turned its nose up at a golden opportunity several years earlier. In 2002, after a Defense Humint officer made an approach through an intermediary, a fiery Islamist leader named Sheikh Hassan Dahir Aweys let it be known that he was open to establishing a relationship with U.S. intelligence. Aweys would go on to become the spiritual adviser to the Islamic Courts Union and al-Shabaab, a man in regular contact with the United States’ highest-priority targets in the Horn. “All Aweys wanted,” said a special mission unit officer, was “respect.” But in 2002 mid-level Defense Humint managers had “no interest” in pursuing a relationship with him, the officer said. “It wasn’t sexy.… Aweys was a nobody. Few of us believed he was destined for greatness, loosely defined.”
The fact that the U.S. government had already designated him a supporter of terrorism didn’t help. “They didn’t want us meeting with an actual terrorist,” said the officer, many years later. “It was still early in the game. Nowadays nobody would think twice about it.”
As the situation in Somalia worsened, a human intelligence source with the access and placement of Aweys would have been invaluable. “We could have been in his camp in 2002,” said the special mission unit officer. “It would have been a lot of work for an unknown return, but looking back through a better lens, we probably should have [done it].… That was a missed opportunity.”
* * *
With the AC-130s gone, JSOC turned to the Navy when it needed to strike high-value targets. On March 3, 2008, the task force tracked Saleh Ali Saleh Nabhan, a twenty-eight-year-old senior East Africa Al Qaeda figure, to a compound in Dhobley in southwestern Somalia. JSOC had been on Nabhan’s tail for five years. Now they had him in their sights. There then followed what the senior intelligence official described as an “unbelievably painful” decision-making process as JSOC shoved its request to strike Nabhan through the upper layers of the U.S. government. Finally, President Bush signed off. A Navy vessel fired two Tomahawk cruise missiles at the compound, destroying much of it. JSOC’s intelligence had been good. Nabhan was indeed at the compound, but he escaped the blasts. McChrystal later complained that the task force made a mistake in only firing two missiles (“to be conservative”) when four would have done the job. “The miss was a bitter lesson for me,” he wrote.
Eight weeks later, there was almost a case of déjà vu for the task force. This time the compound was in Dhusamareb in west-central Somalia, and the target was Ayro, the al-Shabaab leader. Another torturous bureaucratic struggle ensued. “The confidence [the Bush administration] wanted was almost 100 percent, because they didn’t want to have this compound destroyed with a whole bunch of women and children getting lined up,” the senior intelligence official said. That meant the task force had to confirm Ayro’s location as close as possible in time to the missile launch. Flying out of Djibouti, a Chain Shot aircraft—a secret variant of the Navy P-3 Orion sub-hunter—provided real-time video of the compound. President Bush gave his okay and at least four Tomahawks flew across Somalia. Sitting in his headquarters in Balad on May 1, McChrystal watched a screen nervously waiting for the explosions, “worried about the potential impact of a second failed strike on [JSOC]’s standing and its hard-won freedom of action.” But this time, JSOC made no mistake. The missiles devastated the compound, hitting at about 3 A.M. and killing Ayro and several colleagues. JSOC’s “freedom of action” was safe, but it would have to wait for its reckoning with Nabhan.29