by Steve Coll
President Clinton approved a limited “proof of concept” mission to launch Predators over Afghanistan in September. The concept to be proven, recalled Air Force Secretary Whit Peters, was the CIA’s ability to fly the Predator “from barren and difficult airfields, controlled via satellites from a ground site many thousands of miles away.” The CIA would complete the mission without its pilots or commanding officers ever leaving the Virginia suburbs.14
The bin Laden unit drew up maps and plans for fifteen Predator flights, each lasting just under twenty-four hours. They decided to fly over places they had previously identified as bin Laden’s main haunts, especially in eastern and southern Afghanistan. They also lit up their agent network on the ground. They sought detailed reporting about bin Laden’s movements, hoping to steer the Predator overhead and photograph him. Clarke urged the White House to be prepared to attack bin Laden if the Predator found him. Berger cautioned that they would need more than just bin Laden’s location—they would also want a reliable forecast of his plans or movements during any cruise missile flight times.
Previous operations in the Balkans and Iraq had shown that the Predator was most effective in daylight hours. The drone could carry night vision equipment, but the images were much harder to decipher. Daylight hours in Afghanistan began in the dead of night in Virginia. A large video screen loomed in the middle of the CIA’s makeshift flight operations center. Air Force drone pilots and CIA officers from the Counterterrorist Center and the bin Laden unit huddled in the darkened room on the wooded Langley campus from midnight to dawn, watching black-and-white aerials of Afghanistan unfurl eerily before them.
Richard Clarke would drive out after midnight, clear the CIA’s security gates, park in the darkened parking lots, and wander through empty hallways to the flight center. Other curious visitors arrived at odd hours as well. They were like a secret society of video game junkies, role-players in a futuristic scenario, and they were well aware of their role in pioneering a kind of technical espionage that Hollywood might promote. They sipped coffee and talked to their pilot. “Oh, look at that truck! That truck looks like the one he uses! Follow that truck!” Remembered one participant: “It was very much the O.J. thing, with a helicopter following a car down the freeway.” Clarke wrote Berger that the images were “truly astonishing.” Berger replied with encouragement, but also cautioned: “Unfortunately, the light at the end of the tunnel is another tunnel.”15
The Taliban’s air defense units monitored flights across the Uzbek border. One night the CIA’s drone flew above a Taliban airfield where a MiG fighter jet prepared to take off on an intercept mission. In the Langley operations room they could see the fireball light up the MiG’s tail as it thundered down the runway. The Predator’s eavesdropping equipment captured chatter between the MiG pilot and the control tower. “I can’t find it! There’s nothing here!” the Taliban pilot complained to his commander. Suddenly the Predator’s camera picked up the MiG flying right toward the drone at jet speed. “As the MiG flew by, half the people in the room ducked,” recalled an American official who was watching from Langley. The MiG pilot never spotted the drone and returned to base. At Langley the audience slumped in its chairs, relieved and amazed.16
While hovering over Tarnak Farm outside of Kandahar, the Predator photographed a man who appeared to be bin Laden. An agent reporting from Kandahar suggested that the Saudi had come to visit one of his wives. The camera showed a tall man in Arab robes surrounded by armed bodyguards walking from a building previously mapped by the CIA as bin Laden’s residence to a tiny mud-brick mosque across the way. There was no way to be 100 percent certain that the man was bin Laden, but the evidence was very strong. On two other missions the Predator recorded images of a man who CIA analysts later concluded was probably bin Laden, but in these cases they were less certain than they were about the Tarnak case.17
Their arguments about the mission continued even as the Predator flew. One issue was security and secrecy. As Taliban radar tracked the flights, some at the CIA worried about Uzbekistan’s exposure. They did not want to jeopardize their work with the Uzbek commando unit. A downed Predator would also be a propaganda coup for the Taliban. The drone carried little sensitive equipment—most of its sophisticated electronics were housed in the pilot’s remote console. Yet nobody wanted a Predator to be captured, and CIA officers sometimes felt that the Pentagon overestimated the drone’s ability to hide from enemy aircraft and ground fire. Richard Clarke discounted the strength of the Taliban air force: Its pilots never fired the few air-to-air missiles they carried on their MiGs, and if they tried, they would probably just blow themselves up, he said. He badgered the CIA not to worry so much about Predator accidents. “The pilot will return safely to base,” he noted sarcastically.18
Their fights about money were even more pointed. When one Predator crashed on takeoff, the Air Force tried to bill the CIA for the replacement cost. Tenet, Pavitt, and Black protested. They had not budgeted money for broken $3 million drones. Aggravated, the Pentagon’s officials battled back. Whit Peters at the Air Force felt that the CIA’s managers wanted “to run everything and pay for nothing,” as he recalled it. “They like to have sexy toys that do interesting things so they can claim credit … and of course, they don’t want to pay for it.” For their part the CIA’s officers felt they were pushing the Pentagon to innovate. Left to its own devices the Air Force would bury the Predator’s development in excruciating testing schedules, reams of written specifications, and elaborate contracts. The CIA could move much faster, the agency’s officers felt. The Air Force ought to pay for the Afghan operation, CIA officers believed, in part because the Pentagon was learning more about the drone’s capabilities in a month than they could in half a year of sterile testing in Nevada. Memos and emails ricocheted around Virginia and back and forth to the White House, but still the funding question went unresolved.19
By mid-October fierce winds gathered in northern Afghanistan. On some flights the Predator’s meek engine had trouble propelling the drone across the mountains. The Predator kept drifting back toward Uzbekistan. Temperatures plummeted, and wing icing became a more worrisome problem. They knew from Balkans experience that the Predator was a very difficult plane to fly in bad weather. The White House and the Counterterrorist Center halted the operation. The Afghan mission had always been designed as a finite experiment.20
During the winter hiatus Black and others at the CIA hoped the lawyers would resolve the treaty questions that had postponed testing of an armed version of the Predator. Having seen the images of bin Laden walking toward the mosque at Tarnak, Black was now a vocal advocate of affixing missiles to the drone. Here was the clean shot they had been seeking for more than two years: positive identification of their target, no questionable human agents, no delay.
At the White House and the Pentagon, too, those involved hoped to be flying Predators again in the spring—if they could find the money.
THE DRONE IMAGERY had brought them back once again to Tarnak Farm on the sagebrush-strewn desert flats outside of Kandahar. Tarnak had been the target of the CIA’s first secret plan to kidnap bin Laden, back in 1998. More than two years later the United States, an unchallenged global power with a military larger than all of its serious rivals combined, with aircraft carrier groups and B-2 bomber wings that could strike any target worldwide in twenty-four hours or less, still found itself stymied by this lightly defended mud-walled compound of several hundred acres, a fort that would not even have intimidated horsebacked Pashtun raiders several centuries before. Tarnak’s water-streaked concrete office building—the onetime agricultural extension office of a doomed Afghan communist government—peeked over an empty plain that could be crossed from all directions. There were no mountains within miles, no rock walls, no gorges, no natural defenses of any kind. Yet Tarnak flummoxed Clinton and his closest national security advisers. To a great extent the problem was one of foreign policy: As Massoud’s intelligence aides put it, the Ame
ricans insisted on capturing the king without disturbing the pawns. By refusing to declare the Taliban an enemy Clinton and his Cabinet made Tarnak a very complicated target. In another sense, however, the farm was a symbol of the political-military problem now commonly referred to in Washington as “asymmetric warfare,” which described the advantages that terrorists and guerrillas can exploit against a superpower by virtue of being small, dispersed, and blended with civilian populations.
Clinton’s national security and intelligence team spent many hours studying satellite photographs of Tarnak’s flat-roofed, one-story residential buildings, clustered in several tiny villages behind the compound walls. At the Pentagon, targeters with the Joint Chiefs of Staff crunched trigonometry equations and blast calculations to determine which of Tarnak’s little concrete boxes—no more than sheds, by American standards—would collapse on their inhabitants if one or two or three cruise missiles slammed into the particular house where bin Laden slept. One of the nearby sheds was a mosque. Another was a medical clinic. American military doctrine presumed the sanctity of such buildings. This was the purpose of the Pentagon’s missile math: to determine which available munitions would be most likely to destroy the Tarnak house where bin Laden stayed while knocking down the fewest neighboring houses. Alone among the world’s militaries, the United States had the capacity to ask and answer such questions. It was also the first military power in world history whose leaders argued day after day in conference rooms about the mathematical nuances of their destructive power.21
Then there was the child’s swing. Families lived at Tarnak. The CIA estimated that the compound contained about one hundred women and children—bin Laden’s family and family members of some top aides. There were laundry lines, and agent reporting and satellite imagery clearly showed a wooden swing near some of the residential buildings. There were no pictures of any kids actually swinging, but the children were officially presumed to be nearby.22
The swing made an impression on Clinton. The president recognized that his conflict with bin Laden was multidimensional. The propaganda war mattered. “It’s almost like he was daring me to kill them,” Clinton recalled of the women and children at Tarnak. He had learned through hard experience: “I do not care how precise your bombs and your weapons are, when you set them off, innocent people will die.”23
Tarnak was now the visual locus of their elusive enemy. The Predator image of bin Laden in his flowing robes at the farm compound was copied onto videotape by the CIA. It was a startling loop, convincing and ominous. Tenet brought the tape to the White House and played it for Berger and Clinton. The video’s eerie power seemed to convert Tenet to the Predator’s cause. He carried the video to classified briefings on Capitol Hill and raved about the drone’s achievements. They were getting closer to their mark, he hoped. Clinton, too, was encouraged by the Predator experiment. The president remained interested in the possibility of a Special Forces raid in Afghanistan against bin Laden. But the Pentagon and CIA’s “strong and constant view,” as Clinton recalled it years later, was that such operations were likely to fail without better intelligence and a great deal of lead time. The Predator images were intriguing, but they did not provide enough.24
AS THE PREDATOR FLEW above him, bin Laden pressed his two-front war below, against Massoud and the United States.
In September, al Qaeda’s jihadist volunteers in Brigade 55, based at Rishikor, a former Afghan army camp on Kabul’s southern outskirts, joined the Taliban’s late-summer thrust against the Northern Alliance. The CIA estimated al Qaeda’s annual budget at $30 million, much of it spent on the Taliban and war-fighting operations in Afghanistan. Thousands of Pakistani madrassa students, aided by ISI, joined Taliban forces on the outskirts of Taloqan, the ramshackle northern town that now served as Massoud’s headquarters. Loaded with cash, they bribed Northern Alliance commanders to switch sides. Aided by unusually precise artillery fire—a bombardment that some American analysts interpreted as evidence of direct participation by Pakistani army officers—they stormed the town and sent Massoud and his men reeling into nearby Badakhshan province. Suddenly Massoud faced the loss of his overland supply lines to Tajikistan. It might take another summer of fighting for the Taliban to cut him off completely, but if they did, Massoud would have to either seek exile in Dushanbe or bottle himself up in the Panjshir, living off what he could capture and forage. The Taliban might be weakening politically among Pashtuns, but its resources—money for bribes, ammunition, and vehicles; volunteers from abroad; expert military advice from Pakistan—did not slacken.25
A month after Taloqan’s fall, on October 12, a small tender boat packed with explosives glided alongside a 505-foot, American Arleigh-Burke class guided-missile destroyer docked at Aden, Yemen. The USS Cole was a billion-dollar command and attack ship equipped with computer-linked radar that could follow more than one hundred airplanes, ships, and missile targets at once. It had relatively little defense, however, against three suicide bombers in a thousand-dollar skiff. The attackers blew a hole twenty feet high and forty feet wide in the Cole’s hull, killed seventeen American sailors, and wounded thirty more. With just slightly more skilled execution, CIA analysts later concluded, the bombers would have killed three hundred and sent the destroyer to the bottom.26
There had been no specific tactical warning that the Cole was a target. The CIA had circulated a classified analysis the day before the attack that highlighted the growing al Qaeda threat in the region, but it provided no specific warning about the Cole. A Pentagon intelligence analyst resigned on October 13, declaring that his warnings about al Qaeda in the region had been ignored and suppressed by his superiors. None of his analysis involved specific threats against the Cole, however. Daniel Benjamin and Steven Simon, the former White House counterterrorism aides in Richard Clarke’s office, who had left government by the time of the Cole attack, later accused the U.S. Navy of blatantly ignoring the al Qaeda threat. “A more telling display of the persistent disbelief” that bin Laden and his network posed a danger “would be hard to imagine,” they wrote. They also blamed Anthony Zinni, the regional commander in chief, for permitting refueling operations in Yemen. Zinni defended his Yemen policy with arguments similar to those he called upon to advocate American engagement with General Musharraf in Pakistan. Even where Arab and Muslim governments were highly imperfect, Zinni argued, it was in America’s best interests to deepen contacts and alliances despite the risks.27
The Cole attack hit officers and analysts in the CIA’s Counterterrorist Center very hard. The millennium period had been a succession of terrifying near misses, but they had gotten through unscathed. Now they had taken the first big loss at bin Laden’s hands since the Africa embassy attacks. In the initial weeks the center was consumed by searches for evidence about the attackers and their links to bin Laden. They found connections between the bombers and an al Qaeda operative who had recently spent time at a Kandahar guest house. But they could not prove bin Laden’s personal responsibility for the attack—at least, the evidence would not meet the standards of a criminal indictment. Nor could they provide specific proof of bin Laden’s role that Clinton could cite if he wished to publicly justify retaliation. Yet the CIA’s officers told colleagues that they were dead certain of bin Laden’s involvement.28
“We’ve got to change the rules,” the CIA’s bin Laden unit chief argued in the aftermath. It was time for the agency to try to break the policy stalemate about the Taliban. Al Qaeda was growing, and its sanctuary in Afghanistan allowed ever more ambitious operations. Within the CIA and at interagency White House sessions the Counterterrorist Center officers spoke starkly. “Al Qaeda is training and planning in Afghanistan, and their goal is to destroy the United States,” they declared, as one official recalled it. “Unless we attack their safe haven, they are going to get continually stronger and stronger.”29
Clarke was the only senior White House official who agreed. Clinton would be president of the United States for just th
ree more months. His vice president, Al Gore, from whom Clinton had grown estranged, was locked in a close election campaign against the Republican governor of Texas, George W. Bush. Any military attack Clinton launched now would rebound on Gore one way or another. If the president fired at bin Laden and missed or if he killed Arab or Afghan women and children, he risked making the White House appear reckless or incompetent on the eve of the national vote. Undoubtedly Clinton would be accused by talk show conservatives, however absurdly, of launching the strike to boost Gore’s chances. In any event, few of Clinton’s senior national security aides supported a retaliatory attack. Even after the Cole bombing, Clarke could not persuade Defense Secretary William Cohen or his top uniformed officer, Hugh Shelton, to take an offensive strike against al Qaeda or the Taliban seriously. “Although we fully shared Mr. Clarke’s anger and frustration,” recalled Madeleine Albright, “it was not clear that air strikes directed at training camps would cause any significant disruption to al Qaeda.” Shelton produced a paper after the attack describing thirteen options for the use of American military forces in Afghanistan, including several plans to conduct Special Forces raids to capture or kill bin Laden. Shelton’s chief of operations later described the paper as essentially a primer designed to “educate” Sandy Berger and aides such as Clarke about the “extraordinary complexity” of actually going ahead with any of the options. Clarke had by now given up on the Pentagon. Their “overwhelming message,” he said later, “was ‘We don’t want to do this.’ ” Even after a direct assault on American sailors aboard the Cole, the consensus among the Pentagon’s civilian and uniformed leaders, Clarke remembered, was “that their capacity not be utilized for commando operations in Afghanistan.” That left the CIA and the possibility of using Massoud’s Northern Alliance as a proxy force to attack al Qaeda. Clarke had set aside his earlier skepticism about Massoud and now agreed on the need for infusions of guns and money. He encouraged Black and Rich, the bin Laden unit chief, to go ahead with a new Afghan plan.30