The Wizards of Langley

by Jeffrey T Richelson

  Once Iraqi troops began moving, six to seven ground forces analysts were selected to cover Iraq. Their shifts increased to fourteen to sixteen hours a day. Not surprisingly, people grew tired and made mistakes, including incorrectly entering the coordinates of SA-6 antiaircraft missile sites into the database.29 Meanwhile, back at Building 213, the interpreters were being overwhelmed by the flood of incoming imagery. Further complicating matters was the lack of sufficient broad-band transmission capability to send much of the imagery to Central Command (CENTCOM) headquarters in Riyadh electronically. Thus, a message based on KH-11 imagery would normally take over an hour to get to the field, and an actual image would take between four and fourteen hours to arrive. Messages based on radar imagery would take two to three hours, whereas images would take between six and twenty-four hours. Those images that took the longest to arrive might arrive via aircraft rather than the airwaves. An Air Force jet was dispatched each evening carrying overhead images.30

  The war also revealed differences in the interpretation cultures at NPIC and the theater with respect to bomb-damage assessment. NPIC judged aircraft or tanks to be destroyed if destruction was clearly shown by satellite or aerial photography, whereas CENTCOM analysts factored in pilot reports in assessing the impact of bombing raids. In addition, many of the small holes on the outside of a target, which meant serious internal damage, were not detectable—even by high-resolution U.S. imagery satellites. As a result, CENTCOM concluded that 1,400 of the 4,280 Iraqi tanks believed to be deployed in Kuwait were destroyed, but NPIC could confirm only 358 as destroyed.31

  THE END OF PROGRAM B

  The late 1980s ushered in the beginning of what eventually would be a wide-ranging restructuring of the NRO, a restructuring that would have significant implications for the CIA’s Directorate of Science and Technology. In a November 1988 letter to Senator David Boren, chairman of the Senate Select Committee on Intelligence, NRO director Pete Aldridge observed:

  As you are aware, over the last year and a half we have conducted an extensive study of the organizational structure of the NRO. While I am convinced that the NRO is extremely effective and responsive to the many needs of the national intelligence community, the dramatically expanding collection requirements, the increasing technical complexity of the targets, the constrained budgets, and the growing diversity of the operational users demand that the NRO become even more effective and efficient.32

  Aldridge reported that he had discussed the study’s recommendations with Secretary of Defense Richard Cheney and Director of Central Intelligence William Webster and he was directing the development of plans to implement the recommendations. Specific changes would include the creation of a centralized systems analysis function “to conduct cross-system trades and simulations within the NRO”; creation of a “User Support” function to improve NRO support to intelligence community users as well as to the growing number of operational military users; and the dispersal of the NRO staff to the new units, with the staff being replaced by a group of policy advisers. In addition, Aldridge foresaw the establishment of an interim facility “to house the buildup of the new functions and senior management.” The ultimate goal, projected for the 1991–1992 period, would be the “collocation of all NRO elements . . . in the Washington, D.C., area.”33

  Not planned, and Aldridge pointed this out in his letter, was any change in the nature of Programs A, B, and C as “distinct elements” of the NRO. A study of the NRO conducted by former OSO director Barry Kelly and Rear Admiral Robert K. Geiger, former head of the NRO’s Program C, and completed in July 1989 had not recommended any change in that twenty-year-old arrangement.34

  But they did recommend collocation as a means of increasing the DNRO’s authority—as well as installing the Director of OD&E as the head of Program B. They noted that “because the DDS&T reports directly to the DCI, there are real and perceptual problems regarding his willingness to support a DNRO decision that is unfavorable to the CIA, or to appeal it with the DNRO. Instead, the DDS&T can . . . take the issue directly to the DNRO’s boss, the DCI.” In addition, they noted that “the proximity of the DDS&T to the DCI also tends to cause the DCI to look first to the DDS&T for support regarding NRO issues.” Finally, they observed that “the DDS&T is limited by other responsibilities and can spend only about 20 percent of his time on NRO and Program B matters,” but that “the effective management of . . . Program B requires . . . someone who spends the majority of his time working Program B and NRO issues.” 35

  In July 1989, Webster and Cheney wrote Boren a joint letter reporting on further plans for restructuring NRO’s internal operations. The proposed changes, which were largely based on the Geiger-Kelly study, would include establishing a planning, analysis, and evaluation capability within the NRO to support program decisions; creating a deputy director for military support; and formally designating the CIA’s Director of the Office of Development and Engineering as Director of Program B in order “to provide a full-time manager for Program B.”36

  Cheney and Webster also informed Boren that the Program A, B, C structure would remain in place:

  [W]e reaffirm our previous conviction, supported by the DNRO’s current reassessment, that a business-line structure, that would attempt to give each Program Office the responsibility for a unique mission area, is neither a viable or effective restructure alternative. We want to preserve a beneficial degree of competition between the Program Offices, as appropriate to a problem. Competition is also vital to sustaining the motivation of the Program Offices and our ability to develop creative solutions to intelligence requirements.37

  However, the policy of maintaining the traditional structure would not last through the Bush administration. In the late 1980s, Congress, particularly the Senate Select Committee on Intelligence and chairman David Boren, had suggested the need for NRO to reorganize and consolidate. The feeling, as described in 1994 by DCI James Woolsey, was that the “NRO . . . was a somewhat decentralized organization, and the various parts of it, from time to time, fell into competition with one another. And that involved, sometimes, competing . . . satellite programs.”38

  Thus, in October 1991, Senator Boren and committee Vice-Chairman Frank Murkowski, in correspondence with the DCI and Secretary of Defense, noted that their committee “recommends reorganization into several directorates and collocation of major NRO elements as expeditiously as possible.” The proposal was approved by Robert Gates, who had replaced Webster as DCI in 1992, Cheney, and Bush.39 The President formally ordered the restructuring in National Security Directive 67, “Intelligence Capabilities, 1992–2005,” which he signed in late March.40

  By the time Bush signed the directive, he had been apprised of the suggestions of a task force chaired by former Lockheed CEO Robert Fuhrman. Its members included four former senior intelligence community officials—Lt. Gen. Lincoln Faurer (NSA), Lt. Gen. Edward J. Heinz (Intelligence Community Staff director), Maj. Gen. Ralph Jacobson (Program A), and Evan Hineman—and two serving intelligence officials, John P. Devine, the NSA’s Deputy Director for Research and Engineering, and NPIC director Leo Hazelwood.41

  Among the panel’s recommendations was the termination of the Program A, B, C structure. The panel observed that the traditional organizational arrangement did not “enhance mission effectiveness.” Rather, it had led to “counterproductive competition,” which “makes it more difficult to foster loyalty and to maintain the focus on the NRO mission.” They recommended, “in order to foster an improved NRO corporate spirit, and to better serve the intelligence needs of the nation,” that the NRO be organized around the imagery and SIGINT disciplines. The panel noted that “such a restructure will lessen competition between NRO program offices as a driving force for creativity,” but believed that the NRO director would be able to find “other and more effective ways of eliciting the most creative and effective ideas for meeting the nation’s intelligence needs.”42

  Gates publicly announced the restructuring befo
re a joint public hearing of the Senate and House intelligence oversight committees in April 1992, at a time when the NRO’s existence was still officially secret. Gates told his audience that there would be a “far-reaching internal restructuring of the Intelligence Community organization responsible for designing, building, and operating our overhead reconnaissance assets.”43

  The restructuring that Gates referred to took effect the following month. Initially, imagery activities were assigned to the CIA’s Office of Development and Engineering, and SIGINT became the responsibility of the Air Force Office of Special Projects. Two tasks that remained were the transfer of all 700 Program A personnel from Los Angeles to Washington—a prospect that perturbed some southern California congressmen and no doubt many of Program A’s staff—and the creation of NRO imagery and SIGINT directorates that were purely NRO entities (and that fully integrated Air Force and CIA personnel, along with personnel from the Navy, NSA, DIA, and other organizations).44

  A major question not yet resolved in late 1992 concerned the impact of the restructuring on the Office of Development and Engineering. It was clear that it would continue to be the source of personnel to work on NRO programs. And in accord with McMahon’s instructions to Kohler, it had often functioned more as a component of the NRO than of the CIA. The distinction was emphasized by its having its own building, as well as its own career service—distinct not just from the agency’s but from that of the rest of the science and technology directorate. But the restructuring meant that eventually there would be no more CIA/Program B satellite programs, only NRO programs that CIA personnel participated in—which had been Charyk’s and McMillan’s vision three decades earlier.

  BACK TO THE FUTURE

  In August 1993, James Hirsch, faced with budgetary and personnel downsizing, effectively reversed one of his and Evan Hineman’s changes, when he established the Office of Technical Collection (OTC). OTC absorbed, in their entirety, the Office of SIGINT Operations and the Office of Special Projects. It also took charge of some of the projects that were assigned to the Office of Technical Service, including some of its clandestine imaging activities.45 Since the OSP-managed activities had largely been the responsibility of OSO until the special projects office was established in 1987, in one sense the clock had been turned back.

  The merger returned responsibility to a single office for the CHESTNUT sites, the Special Collection Service operations, and emplaced sensors—and once again assigned one office responsibility for both SIGINT and MASINT collection. But as its name implied, OTC gave the two activities more equal standing. Indeed, the first head of OTC was Peter Daniher, who had headed OSP and was chosen over OSO head Joseph B. Castillo Jr. to run the new office.46

  Daniher had served in the intelligence directorate’s Office of Scientific and Weapons Research (established in 1980 by merging OSI and OWI) and had ambitions to be the Deputy Director for Intelligence. When it became clear to him that he was unlikely to achieve that position, he shifted his focus to the DS&T and became director of the special projects office in October 1989 when Gary W. Goodrich moved up to become Hirsch’s deputy.47

  Hirsch had a number of reasons for establishing OTC in addition to resource pressures. The seemingly disparate activities had some significant common aspects. The data they produced could all be processed using digital techniques, and Hirsch believed it was important that any advances be shared. Developing access to sites to place sensor equipment, whether clandestine SIGINT, clandestine imaging, or clandestine MASINT, involved common problems and challenges. There was also commonality in development and testing. In Hirsch’s view, there was no good, logical way to divide the activities perfectly, and that what was important was proper development and testing, irrespective of whether the system was electronic, optical, or chemical.48

  THE WALKING STICK

  In late 1992, President Bush authorized Operation Restore Hope—the deployment of 25,000 U.S. troops to Somalia to ensure that U.N. food, medicine, and other supplies were delivered to those in the war-torn country in need of such assistance. By May 1993, the supplies were reaching their targets, famine was receding, and the nation was relatively peaceful. Most U.S. troops were withdrawn, and Somalia was turned over to a U.N. peacekeeping force.49

  Then, in a move supported by the Clinton administration, the U.N. mandate was expanded to include the rehabilitation of Somali political institutions as well as the nation’s economy. The action infuriated Gen. Mohammed Farah Aideed, a warlord whose Somali National Alliance (SNA) had become the primary power in Mogadishu, the Somali capital. In early June, Pakistani peacekeepers were ambushed and killed after inspecting Aideed’s radio transmission center. Not long after, the U.N. senior representative in Somalia, Adm. Jonathan Howe, issued an arrest warrant for Aideed and offered a $25,000 reward for his capture.50

  The next month, the situation took a turn for the worse, when a U.S. component of the peacekeeping force, the Army’s 10th Mountain Division, killed between twenty and fifty Aideed aides and operatives in an attack on the SNA’s command center. The attack not only failed to achieve its objective of eliminating Aideed and the SNA as an obstacle to the U.N. mission but also led the warlord to declare war on the U.N. forces.51

  Finding and capturing Aideed became a top priority for the CIA in Somalia. Trying to locate the warlord through intercepted radio traffic proved impossible in Mogadishu’s “pre-electronic state.”52 In early October 1993, an article in the Washington Post recounted how intelligence problems, including a lack of intelligence exchange between the nations involved in the peacekeeping effort, hampered the search for the elusive warlord.53

  One problem not mentioned was the fatal flaw of a CIA asset, which possibly prevented his case officer from using him to plant a homing beacon on Aideed. OTS had implanted such a beacon in an ivory-handled walking stick. The asset, a minor warlord, was to give the stick to Aideed as a symbol of friendship. As long as Aideed carried the stick with him, the CIA would be able to find him.54

  But in late August, before the CIA’s man could deliver the stick to Aideed, he played Russian roulette once too often. By late September, with pressure building to produce a success for the Clinton administration, the deceased warlord’s case officer was told that another troublemaker might be captured. The potential target was Osman Ato, a wealthy businessman, arms importer, and Aideed financial supporter. Another CIA contact was willing to deliver Ato for the right amount of money.55

  Cane in hand, the contact was soon climbing into a car that would take him to Ato. Helicopters tracked the car, using the stick’s beacon, on its winding ride through north Mogadishu. When the car stopped for gas, an operative on the ground reported that Ato was in the car. Commandos from the Delta Force, which had arrived in late August expecting to capture Aideed, were launched in pursuit of their new target.56

  Minutes later, a helicopter descended, and a sniper leaned out and placed three shots into the car’s engine block. As the car came to a stop, the commandos slid down ropes dangling from Blackhawk helicopters, surrounded the car, and put Ato in handcuffs. An hour later, Delta’s commander, Maj. Gen. Walter F. Garrison, had a big grin on his face and was telling CIA officer Garrett Jones that “I like this cane.”57

  SPYING ON CANCER

  On October 11, 1994, James Woolsey, the Clinton administration’s first DCI, appeared at a Capitol Hill briefing. He was not there to discuss terrorism, arms proliferation, drug trafficking, or the threat from rogue states. At the request of Susan Blumenthal, the deputy assistant secretary for women’s health at the Department of Health and Human Services, he was attending a briefing on breast cancer. He was not there as an observer but as the most senior representative of a CIA effort to make some of its technologies available to fight the disease. Woolsey told the audience that “what we are trying to do is to use the tools developed to protect the lives of 250,000,000 Americans [to] help save the lives of a substantial share of the 46,000 women who die annually of breast cancer.�
��58

  Earlier in the year, Blumenthal had written to other government agencies inquiring whether they had technologies that might contribute to President Clinton’s initiative to fight breast cancer. In July, ORD invited the National Information Display Laboratory (NIDL), an organization sponsored by the intelligence community to help bring advanced commercial and consumer technologies into the government, to prepare a background paper for Blu-menthal. The paper described some of the advanced image-processing techniques that might be applied to the field of mammography.59

  The Defense Intelligence Agency digitized some sample mammography for NIDL experimentation, which proved sufficiently interesting that Blumenthal and a committee of medical experts visited the laboratory in September 1994. The image-processing technique demonstrated to the group led Blumenthal to invite the laboratory to participate in the Capitol Hill symposium. For that meeting, NIDL, with the assistance of the Community Management Staff and ORD, set up several demonstrations of advanced, but unclassified, image-processing techniques that were applicable to medical procedures.60

  The techniques included using experimental high-definition television and head-mounted displays to examine digital images, pattern recognition tools, and a three-dimensional system for detecting lesions in magnetic resonance breast scans. The briefing also included a demonstration of serial change detection—the subtraction of the images in one photo from a later photo to identify new objects—applied to serial mammo-grams.61 Such a system had been developed to enable NPIC to identify new missile sites or other new military developments at a target area.