A lot had been learned in a year and a half of flying over Afghanistan, the most important lesson being that when there was a battle on—“troop in contact”—everything else was secondary and everyone wanted pictures. Though Global Hawk was intended to fly a route planned in advance and to take a set of specific images based upon formal collection requirements, in actual operations it ended up being used as much for ad hoc tasking, often deviating 100 miles or more off its course to put eyes over a developing battle or look at something interesting called in by spotters on the ground. During the Battle of Tora Bora, Global Hawk dropped its planned imagery collection profile altogether and started tracking Taliban and al Qaeda positions and cave entrances, either using its infrared camera to detect campfires or receiving tip-offs to take a closer look with its sensors into crevices and cracks off angle from satellites. Aircraft and AC-130 gunships in the area would undertake airstrikes, which Global Hawk could then instantly confirm via star-shaped infrared flashes, recording bomb explosions, the sparkle of success.
Such instant gratification made quite a contrast with doctrine and even practice before the 2003 war: in old-fashioned war, the army corps commander, in charge of three to five divisions, would rely upon his own intelligence units—either field artillery radars or organic collection assets—out from the front. A fire support team would determine the best way to attack beyond the range of the divisions. If airstrikes were desired, a liaison would nominate the target through higher headquarters to the daily targeting board at the air command center, which would then task flying squadrons, which would fly the requested missions anywhere from one to three days later. If the target was mobile, the nominated target would have to be meticulously tracked and its position updated to the air guys as many as three times a day. The obvious question then asked in the fall of 2002 was “When we find a target, why not just kill it right then?” Capabilities and communications had certainly improved; what was needed was closer cooperation at the working level and a change in practice to reflect the new capabilities of the Data Machine.
In Iraq, a beefed-up air force support unit arrived at the army corps headquarters to figure out exactly how to deliver instant support, and it became fully integrated into the command post. Ground forces were planning to move rapidly and skirt most Iraqi defenses, but ahead of the pack were special forces and long-range reconnaissance elements stealthily operating in areas where attacks needed to take extra care. So in addition to the normal concerns about civilian casualties and collateral harm, immediate attacks on midrange targets of importance to the ground force needed to be carefully cleared to avoid friendly fire. The rules were that either a forward controller (a JTAC) or a pilot had to see a target or that the intelligence needed to be confirmed by two sources in situations where neither a JTAC nor a pilot could see it, the latter turning out to be the case in more than 90 percent of the attacks that would take place in March and April 2003.3
The Rand Corporation would later write of the 2003 war: “U.S. forces encountered little resistance from the Iraqi Army during the invasion.”4 And despite the sandstorm, at the crucial moment of seeming crisis for the offensive, the Machine didn’t stumble. With hundreds of embedded reporters scrutinizing every battalion and company’s moves, it truly looked like the offensive had stalled. When the skies cleared, the press corps heaved a sigh of relief for the 5th Army Corps.
But Grumpy had been overhead all along, nudged north of the thirty-third parallel within fifteen miles of Baghdad before any other airborne intelligence asset, and Grumpy flew over the epic storm as well. It wasn’t just Grumpy. Predators had flown over Baghdad since before the shooting started.5 But once the storm developed, Grumpy dominated, along with two manned airplanes: the JSTARS (Joint Surveillance and Target Attack Radar System), with its gigantic moving target indicator radar that could track vehicles, and the RC-135 Rivet Joint, which collected Iraqi signals and electronic emissions. Grumpy was the only platform that had uninterrupted endurance, and it flew in near space far above the storm, where other aircraft didn’t operate. Peering down, its optical and infrared sensors were blinded by dust, but its synthetic aperture radar was able to see through the muck and provide continuous coverage, reconstructing changes on the ground, particularly movement of armored vehicles.
Silently, or at least with all eyes focused on the men and women on the ground, the Machine went into action: aircraft flying overhead couldn’t see either, but then, all they needed were geographic coordinates to deliver their JDAM weapons.
Using the storm as a shield, Iraqi units east of the Euphrates River changed hide sites and redeployed, constantly being bombed as they did so. Irregular fighters, the Fedayeen Saddam and the Quds Force, flooded south, as did elements of two Republican Guard divisions, moving from near Baghdad to reinforce the Medina Division, which was defending the Euphrates River crossings. They too were constantly bombed.
As the sandstorm dissipated, the 5th Army Corps prepared five simultaneous attacks on Iraqi forces stretching from Lake Razazah in the west to Samawah in the east, the main effort intended to skirt Iraqi defenses to the west and swing around to enter Baghdad through what the US military called the Karbala Gap. However, it looked to a blinded Iraq as if the US force would cross the Euphrates River and mount its main offensive up Highway 8, driving straight for Baghdad. On the morning of March 31, intelligence reports started coming in that the Hammurabi Republican Guard division—equipped with tanks and other armored vehicles—was moving to shore up the defense of that route. Grumpy and JSTARS and Predator and the rest of the sensor pool tracked every move. Within minutes of detecting each moving armored formation, bombs arrived. Fighters and bombers dropped satellite-guided 2,000-pound JDAMs on the Iraqis, pilots plugging in coordinates in the air. Linking directly to the bomber cockpits via chat, the Global Hawk desk provided a “last-look” assessment to confirm that Iraqi tanks were still on the intended aimpoints. Predator, together with shorter-range Hunter drones, sent back immediate bomb damage assessments. Counting tanks, artillery, armored personnel carriers, and wheeled vehicles, the Medina division was reduced in strength from 92 to 29 percent of its equipment and personnel. Forty-eight hours later, the US 3rd Infantry Division was at Saddam International Airport in Baghdad.6
All eyes stayed on the ground as the army and marines blew into the Iraqi capital. The statue of Saddam Hussein was toppled and “Mission accomplished” announced just three weeks later. Everyone on the field of battle, and certainly everyone in the command post, knew that none of it would have been possible without Grumpy and the Machine.7 The army later calculated that of the Iraqi forces in front of the 5th Corps, 421 of 660 tanks were destroyed by air attacks; 423 of 843 artillery guns were hit from the air; overall, 1,144 pieces of equipment, more than half of the Iraqi force facing the US corps, were destroyed in the most lethal air attack ever.8
“We did massive damage to the Iraqi maneuver units to the point that in the interviews later [Iraqi officers]… said they just walked away from their equipment because they knew if they stayed with it it was going to be hit,” General Jumper said at a postwar meeting of air force commanders.9 Compared with Desert Storm, where less than 10 percent of the air weapons expended were precision-guided munitions, 71 percent in the twenty-one-day battle were precision.10 But as even the army said, it was more than just numbers: “The traditional means of summarizing combat effectiveness, and particularly the recitation of gross tonnages of ordnance dropped, are meaningless as a way to measure.”11 In this new form of warfare, it was precision rather than gross tonnage that would provide a clear testament of success. And not just precision, but the individual target now measured in personal-sized ways. The air force declared it the beginning of the age of “mass precision,” claiming that the three Republican Guard divisions were destroyed before the army and marine corps even made “ground contact” with them.12 Major General David Deptula, the air command center operations director, said of Grumpy: “Because we contr
olled it… we could put it where we needed it, when we needed it, and for the duration we needed it.”13 The performance of the ISR platforms during the dust storms, General Jumper added, proved to be “a major turning point” in the war.14
“We,” of course, meant the air force, factual but very unbrotherly in a singular military that later faced the consequences of less money and thus more unmanning. “The war,” of course, would go on for another seven years, and the bigger question—whether the offensive was secured by the Machine or by the troops on the ground—was never resolved. In twenty-one days of fighting, Grumpy provided 3,700 images,15 and Northrop Grumman claimed that the single drone, flying approximately 5 percent of the surveillance missions, accounted for more than 55 percent of the information to facilitate time-sensitive engagements, finding 40 percent of Iraqi armor formations, or 300 tanks in total.16
Of course, there were the usual problems: bugs required recoding. A Predator was shot down on March 28, showing vulnerability and the need for a faster and more robust platform. The 5th Corps’ Hunters, the midsized twin-boomed drones equipped with electrooptical and infrared sensors and able to stay airborne for eighteen hours, provided a soda straw view of the battlefield, pointing to the need for a wider view on this particular drone, akin to another Global Hawk.17 And as for the newer unmanned systems like Raven and Dragon Eye that were seeing some of their first combat? Well, there just weren’t enough of them.
Before the situation in Iraq went all to hell, the Defense Science Board assessed the state of the unmanned, concluding that “little doubt remains as to the operational utility and military worth of UAVs,” particularly for the “all important persistent surveillance of the battlespace.”18 Still, the board pointed out that despite experiences in Afghanistan and their incorporation into the Iraq war, only 175 drones were yet operational.19 The board echoed an earlier Defense Department study that concluded that the Pentagon had spent more than $6 billion on unmanned systems and had fewer than 100 large vehicles to show for it.20 The board tried to settle many of the earlier hesitations of the services in adopting drones—technologically and culturally—pointing out that the loss rates of the Predator and Hunter—even of the ancient Pioneer—were comparable with military and general aviation aircraft per 100,000 flight hours.21
As the board is chartered to do, it also made recommendations for future technologies, some of them quite tantalizing. Bandwidth came first, the board fully recognizing its centrality to the operation of unmanned vehicles and the now-regular practice of reachback. The board also predicted that with the emergence of a single network under the Global Information Grid initiative, there would be “a marked improvement in the available bandwidth.” There still needed to be a much greater integration of sensors with combat troops, whether that came through data links that automatically transferred information between systems, or through ROVER-like black boxes that instantly facilitated the movement of the video and pictures.22 And even with GIG deployment, the board pointed out that movement of intelligence over the “last tactical mile” to remote operators, whether they were navy ships operating in distant waters or special forces teams in the mountains of Afghanistan working at the edge of the network, demanded some kind of communications adjunct. The board recommended greater investment in the use of drones to create its own Internet in the sky.23 And finally, to reduce Predator’s and Global Hawk’s voracious appetites for bandwidth, as well as the increasing manpower demands associated with analyzing the Niagara of data now coming in, the board recommended the development of some kind of “on-board target recognition,” that is, “algorithms to survey large areas and reliably select only targets of interest for transmission.”24 In other words, the drones themselves would look at the video and imagery, sounding an alarm and tipping off analysts and fighters when they detected something of significance. It wasn’t exactly autonomous killer robots, but it certainly had some similarities. The scientists recommended based on cost-effectiveness and supposed logic, the developments all followed without consideration of ethical or policy implications. After all, this is a science board.
The board recommended more of everything, from “fighter-like air vehicles for lethal missions” to “small or micro-UAVs for urban combat.”25 Still, the board observed the chaos associated with a system that had gotten used to building a house without an architect: the system relied on spiral development, the process by which new black boxes and capabilities were added as they became available, and it used developmental prototypes rather than programs of record (which coincidentally allowed piecemeal buying rather than presenting Congress with a total program cost). “There are so many different UAV systems in various stages of development,” the board said, “that they are outstripping the ability to evolve standards and approaches for common mission management.”26 Investments in Predator and Global Hawk were skyrocketing, but the rest of the program was largely parochial and uncoordinated.27 Still, no one anticipated how rapidly and how large the drone force would grow and what its demands (and effects) would be.
Iraqi resistance to the American military occupation stayed “at a low, relatively tolerable level” through the fall of 2003 and even into 2004, Rand said,28 with the Iraqi fighters dismissed as “dead enders” and “former regime elements” by the Bush administration. But by the time Saddam Hussein was captured on December 13, it was clear that the United States didn’t understand the sociopolitical circumstances of the country, including the deep divisions between Sunni, Shia, and Kurd that war and regime change had helped to unleash, as well as the sentiments against this occupation. At the end of March 2004, Iraqis attacked a group of four Blackwater contractor guards in Fallujah, an industrial town west of Baghdad. The four were brutally killed; the crowd burned their bodies beyond recognition, then hanged two from the girders of a main bridge, where citizens celebrated.
The Battle of Fallujah began that November, led by US Marines to retake the city. Elsewhere, the United States battled Shia militia associated with Muqtada al-Sadr. The forces in Iraq increasingly also focused on “force protection”—protecting the troops and their bases from insurgent attack. Presence outside fortified bases was inconsistent and often involved ineffective door-to-door raids. But the real issue for the human presence was supply. Main supply routes from Kuwait and from US-occupied bases became a favorite target of anticoalition forces. The first American casualty from a roadside improvised explosive device occurred in June 2003, and that July, in an announcement of a soldier’s death, the military used the term “IED” for the first time. By the summer of 2004, insurgents began to lay “daisy chains” of roadside bombs (multiple, interconnected weapons) in more-precise attacks. The Rand Corporation later said that the United States had a hard time “recognizing the nature of the problem,” choosing a technology effort to counter IEDs, while ignoring violence and suicide bombers who were increasingly terrorizing the Iraqi civilian population, thus creating more chaos.29
Technology advanced, both on the air force side to handle the high volume of precision-guided weapons that were now being employed—everything needed precise coordinates—and on the ground as well, where the tools like ROVER were opening up eyes to real-time and personal intelligence and a world of black boxes just beyond the reach of the normal soldier.30 The Machine was expanding to the edge.
CHAPTER TWELVE
Flock of Birds
“Who goes in front saves his companion,
Who knows the road protects his friend.”
Let Enkidu go before you,
He knows the journey to the Forest of Cedar.
TABLET III, EPIC OF GILGAMESH
Until the IED became the everything and the only thing for US ground forces in Iraq, the old Pioneer from Desert Storm remained the most ubiquitous drone for ground forces, having flown in Desert Storm in 1991, in Somalia, Haiti, Bosnia, and Kosovo, and again in Iraq, starting in 2003. Pioneer wasn’t the first drone, not even of the modern era, and though two decades after
the first Gulf war, it was still there and dominant,1 no one outside of those who had a vision of airpower rallied behind either unmanned systems or the power of the Machine—that is, until the unexpected shift from big war to civil war, from mechanized protection to guerrilla vulnerabilities, made things so lethal. And then what happened was right out of Genesis: the tribe expanded, begats upon begats raising a nation of unmanned.
Pioneer begat Shadow.2 Pioneer veterans grumbled that Shadow’s flying range was 60 kilometers less than Pioneer’s.3 And whereas Pioneer had to be launched by a rocket-assisted catapult contraption and landed in a large net, Shadow… well, had a similar bulky and complicated launch and recovering process, using arresting gear similar to jets on the deck of an aircraft carrier, demanding a flat, cleared space the size of a soccer field to operate.
But in those twenty years, the technologies had transformed, and everything about the modern drones reduced infant mortality to almost zero. Shadow was lighter, had a more powerful engine that used motor gasoline readily available to ground forces, and could fly 4,000 feet higher than Pioneer and loiter for six hours, almost a third longer than its forefather. The first version of Shadow (referred to as the Shadow 200) was thus a substantial advance in all aspects, and the range didn’t particularly matter because it was no longer just the pioneer, the only drone in the hands of the troops on the ground; it was part of a growing family. Its range, in fact, matched the distance covered by typical army brigade-level operations, the highest echelon to which it was assigned.
The army chose Shadow not just to replace Pioneer but also eventually to replace Hunter; the marine corps shot for an improved Shadow-B with three feet of additional wing to increase fuel storage for greater range and payload to match its tactical needs; and the navy began the search for a vertical-takeoff-and-landing alternative that could operate from ships (initially Fire Scout).
Unmanned: Drones, Data, and the Illusion of Perfect Warfare Page 14
