How the Hi-Tech World of Drones and Spy Craft Is a Lot More Hype Than Reality
The following is an excerpt from Andrew Cockburn's new book, Kill Chain: The Rise of the High-Tech Assassins (Henry Holt and Co. 2015).
On January 2, 2011, the Washington Post reported the imminent deployment of a “revolutionary airborne surveillance system called Gorgon Stare, which will be able to transmit live video images of physical movement across an entire town.” Major General James Poss, the air force’s assistant deputy chief of staff for intelligence, surveillance, and reconnaissance, was quoted as claiming that with the new tool, analysts would no longer have to guess where to point the camera: “Gorgon Stare will be looking at a whole city, so there will be no way for the adversary to know what we’re looking at, and we can see everything.” [Air Force intelligence chief Lt. General] David Deptula was no less effusive, certifying that the system offered “many orders of magnitude” improvement over existing sensors on drones in Afghanistan. . . . Instead of looking at a truck or a house, you can look at an entire village or a small city” with the multiple cameras, simultaneously.
Gorgon Stare was definitely the hit of the year in intelligence-surveillance circles. That October, the U.S. Geospatial Intelligence Foundation honored Sierra Nevada with its 2011 Industry Achievement Award, given annually for “outstanding accomplishments in GEOINT tradecraft.” A year later, Deptula’s successor as air force intelligence chief, Lieutenant General Larry James, was still extolling the system’s wide-area imaging as “very powerful in the [Afghan] battlespace” and relaying further tributes from American commanders in that country. Six months later, the general’s enthusiasm was undiminished. “The combatant commanders love it,” he told an interviewer. Earlier, Air Force Times had highlighted its utility in spotting “squirters,” as people fleeing for their lives in an air attack were popularly known in the drone community. Civil libertarians, no less impressed by the Gorgon’s advertised capabilities, expressed alarm at the possibility that it might be put to use by domestic law enforcement.
First appearing in budget documents in 2008, as a response to Defense Secretary Gates’s insistent request for more surveillance systems, Gorgon Stare, developed and manufactured by Sierra Nevada, essentially consisted of five “electro-optical” TV cameras for daytime and four infrared cameras for night missions. These were mounted on a pod under the right wing of a Reaper drone, while another pod under the left wing processed the images, transmitting them to recipients on the ground and storing them for later retrieval. The intent was for the cameras to provide a four-kilometer-square picture with a six-inch resolution, meaning that a scan of a town would reveal objects as small as six inches. A “chip-out” feature allowed troops on the ground to receive a segment of the overall picture. Indeed, according to its developers, the system would be able to transmit a panorama of sixty-five different pictures to different users, as opposed to the single, narrow, “soda straw” images currently available from drones. Thus a single Gorgon-carrying drone could circle over a town, effortlessly delivering images of selected areas to ground units on request. Not only could the wide area under scrutiny monitor “squirters,” as discussed, it was one more attempt at the dream of being able to look back into the past to discover who planted a bomb. As Deptula himself explained, “You can review it and accomplish forensic study of the area by looking at movement and tracing activity. If you know where an improvised explosive device went off, you can ‘rewind the tapes’ and see where the activity was and what led to it.”
There was one problem. Gorgon Stare didn’t work, a fact of which the air force was perfectly well aware. In the last months of 2010, the system had been subject to an intense program of tests by a specialized air force testing unit, the 53D Wing at Eglin Air Force Base in Florida. The results were damning. The final report deemed the system “not operationally effective” and “not operationally suitable,” breaking down, apart from anything else, on average, 3.7 times every sortie. Officially, Gorgon Stare generated “motion video,” which turned out to be just 2 frames a second (as opposed to “full-motion video” at 24 frames a second). While it was possible to make out cars and other vehicles, it was impossible to distinguish “dismounts” (people) from bushes. One of the test team’s briefing slides that I looked at compared aerial pictures of an air base. One was a Gorgon Stare infrared “full image.” In other words, it showed the widest area of which it was capable. The other came from Google Earth, the free online service available to all. They were identical, revealing buildings and roads, and airfield runways, but nothing smaller and more detailed. Another slide showed a “subview,” a sample of what troops in the field would get if they were to make a request to the drone overhead. It was just possible to make out the cars. People were another matter, merely the faintest of blobs and certainly indistinguishable from bushes.
The bad news continued. The wide-area images, for example, were made up of multiple smaller images taken by individual cameras and stitched together in the processor pod before being transmitted to earth. However, as the test unit reported, the imagery “is subject to gaps between stitching areas which manifests itself [sic] as a large black triangle moving throughout the image.” Not surprisingly, “this causes loss of situational awareness and the inability to track activity when the ‘black triangle’ covers the area of interest.” In addition, the system had difficulty in determining where it was and hence the precise location of any targets it might spot. True, daylight images from the pod that were downloaded when it was on the ground rather than transmitted were clear enough to allow the tracking of individuals “to their point of origin or destination, providing analysis of IED detonations.” But unfortunately, said the report, “GS experiences ‘dropped frames’ during download—making it impossible to track moving targets over that period.” The testing unit strongly recommended that Gorgon Stare not be deployed to Afghanistan.
The air force test unit was not the first informed critic to take a dim view of the vaunted surveillance device. In a withering report on the air force’s request for $78.9 million to spend on Gorgon Stare in 2010, the Senate Armed Services Committee had already suggested that there did not seem to be much point for “moderate-resolution” (i.e., poor-quality) motion imagery and that increasing the camera’s resolution would lead to a “dramatic reduction” in the size of the area that could be covered by one Gorgon-carrying drone, which would consequently require more drone flights, which would make the whole exercise too expensive. Adding insult to injury, the Senate report went on to point out that no one to date had “produced sufficient evidence that forensic analysis of moderate-resolution wide-area motion imagery is productive enough to justify a large investment in sensors and platforms— especially in the absence of effective automated analytic tools.” In plain English, this meant that the idea of using the slow motion video to “rewind the tapes” and unmask the IED layers, as suggested by Deptula, wouldn’t really work, especially as the quantity of imagery would be too vast to be analyzed by humans and so would have to be done by computers (“effective automated analytic tools”) that did not exist. The committee recommended “no funds to continue Gorgon Stare development.”
These harsh verdicts made no difference whatsoever. Gorgon Stare was dispatched to Afghanistan a little over a month after the test report. Safe from prying eyes, it could now bask in uncritical plaudits from General James and others. At the end of 2011, I emailed a marine officer deployed in the battleground of northern Helmand and asked if his experience justified General James’s confidence in the system’s ability to help the troops. After detailing the routine IED injuries inflicted on his unit in the previous four days (one double leg amputation, one foot, one arm below the elbow), he went on: “I’ve never even heard of Gorgon Stare, let alone seen it in use. We’re essentially using the same technology that men used in WWII, Korea, and Vietnam to defeat mine and booby-trap threats—the eyeball and metal detector.”
How could it be that this near-useless system could survive such emphatic rejection not only from the air force’s own testing professionals but from a powerful Senate committee, let alone have little impact where it was needed most? The answer would appear to lie in a classic combination of money, politics, and the ever-diminishing gap between government service and private enterprise, all turbocharged thanks to the Niagara of cash unleashed by 9/11.
In the year before 9/11, Sierra Nevada was a small aerospace company based in Sparks, Nevada, that took in $13 million in government contracts. Twelve years later, that figure had soared to $1 billion, most of it in defense work and much of that “sole source,” that is, without competitive bidding. The owners, Fatih and Eren Ozmen, have long enjoyed amiable relations with politicians. For example, in 2004 they hired Dawn Gibbons, wife of then congressman Jim Gibbons, representing Nevada’s 2nd District, and paid her $35,000 for consulting work. Representative Gibbons was a member of the House Armed Services Committee, in which capacity he helped steer a $2 million defense contract for a helicopter landing system to Dawn’s employer. Everyone denied any connection between the two events. A 2010 Congressional Ethics Office report highlighted the close relationship between Sierra Nevada and PMA, a lobbying firm closely linked to the campaign fundraising efforts of house members powerful in defense matters, including John Murtha of Pennsylvania and Peter Visclosky of Indiana. (PMA dissolved in 2009, following a criminal investigation and the jailing of CEO Paul Maglioccetti, a former senior staffer on the House Defense Appropriations Committee.) Nonetheless, the House Ethics Committee voted to take no action.
It seemed that Sierra Nevada’s owner, Fatih Ozmen, could echo the significant statement of Neal Blue, the CEO of Predator manufacturer General Atomics: “For our size, we possess more significant political capital than you might think.” In fact, there were many points of similarity between these two defense corporations, both so profitably engaged in the booming business of drones and surveillance. One of them was Big Safari, the secretive air force office that awarded and oversaw the Gorgon Stare contract. Charged with overseeing the acquisition and introduction of “special purpose” air force weapons, the office, officially known as the 645 Aeronautical Systems Group, enjoyed a reputation for its ability to cut corners and move programs swiftly without the customary bureaucratic encumbrances. Such prowess was largely due to the authority granted them to award sole-source contracts without the tiresome necessity of allowing firms to compete on price and performance.
A secret air force unit known as Big Safari played a major role in the development and fielding of Predator, partly thanks to the intervention of former air force chief master sergeant Mike Meermans, who had spent the last five of his twenty-two-year service career working on airborne reconnaissance operations on the air staff, where he enjoyed a close friendship with the group’s leadership. Following his retirement, Meermans embarked on a second career as a senior staffer on the House Intelligence Committee, when he had crafted the legislation mandating that total control of the Predator program be assigned to the air force, and that Big Safari be put in charge of its development. Given Meermans’ subsequent third career as vice president for strategic planning at Sierra Nevada’s Washington office, he was presumably already well informed about the company and its programs. Overall, between 2006 and 2013, Big Safari lavished no less than $3.5 billion on Sierra Nevada in sole-source contracts for which no one else was able to compete. As a source coarsely summarized the relationship to journalist Aram Roston, Big Safari and Sierra Nevada “are so close they share rubbers.”
Among those deals was an $18 million contract awarded in 2011 for the Northern Command Sensor Program, a bizarre air force initiative to get a piece of the war-on-drugs action by flying surveillance planes supplied by Sierra Nevada over parts of Mexico and other regions in hopes of gathering intelligence about narco-traffickers and their operations. It is unclear whether the operation yielded any useful intelligence, but it ended tragically when one of the planes flew into a mountain in Colombia, killing three Americans and one Panamanian on board. The pilot of the twin engine Bombardier Dash-8 was blind in one eye and the plane so thoroughly laden with intelligence equipment that the FAA forbade its use for anything other than “crew training and market surveys.” The altitude warning system was broken. It appeared to have been a pointless operation spawned by public-private collusion and mounted on the cheap, a telling counterpoint to the aura of glamorous mystery that envelops the world of hi-tech spy craft and the covert firms and groups that inhabit it.
Excerpted from Andrew Cockburn's new book, Kill Chain: The Rise of the High-Tech Assassins (Henry Holt and Co, 2015). Reprinted here with permission from the author.