The Wizards of Langley

by Jeffrey T Richelson

  At the time Hineman called Huffstutler, NPIC was about eighteen months into the major modernization program required but clearly was not going to meet its 1986 target date. The deputy director told Hufstutler that NPIC was having some problems and he would like him to move into the director’s chair. Huffstutler had served as deputy to Hineman in the weapons intelligence office, they had gotten along, and they thought alike in approaching problems. And after working on Soviet issues for fifteen years, Huffstutler was ready for a change. He was also quite familiar with NPIC’s product, since the strategic affairs and Soviet analysis offices had been among the interpretation center’s biggest customers.30

  In addition to getting the modernization plan back on track, NPIC’s new director had two other concerns—ensuring that NPIC’s products could be used effectively by its customers and straightening out internal procedures. One of Huffstutler’s first acts was to review all of NPIC’s reporting. He concluded that imagery analysts did not really understand their role in the analytic process, that they didn’t quite have the proper feel for how their products were used by analysts in the intelligence directorate. Their reporting did not “separate in consistent and clear ways” what imagery analysts actually saw from their inferences. They might see a group of tanks but state without further explanation that the picture “showed” a military exercise taking place. A reader could not separate what the interpreter actually saw (the tanks) and what he inferred (the exercise) from the picture. Only three days after Huffstutler informed NPIC analysts of his views, their reporting began to change.31

  He also addressed two other problems. One was the standard usage of “NPIC believes” in imagery interpretation reports, which he noted “made everyone mad” because it appeared to allow no room for dissent. He told his analysts they could say anything they wanted that met the standards of professionalism and reporting, but they could not turn their conclusions into NPIC’s conclusions. Huffstutler assured them that others would take their advice more often than not. There were no further complaints about NPIC reporting after that time, according to Huffstutler.32

  He also moved to alleviate a feeling by his imagery interpreters that they were second-class citizens in the intelligence community, an effort that took most of his first year in office. The interpreters were referred to, both by analysts in the intelligence directorate and themselves, as “onesource” analysts. On the analytical totem pole, this seemed to place them under the “all-source” analysts in the intelligence directorate who worked with imagery, SIGINT, human intelligence, and open source data. Huff-stutler argued that it was really a question of different jobs. The imagery analysts had access to data from all sources and used that data in moving from what they saw to what they inferred. They were not simply describing what was in a picture but analyzing the significance of the picture.33

  It took far longer than a year to complete the modernization process—indeed it extended into the terms of Hineman’s and Huffstutler’s successors. Two major aspects to the modernization program were causing the delays. One factor was changes in requirements coming out of the program office. The second was software problems.34

  The program itself involved new work stations, new data bases, new connectivity to facilities (such as DIA and field activities), and new measurement equipment. Other steps needed were to put the Priority Exploitation Group at Ft. Belvoir on a twenty-four-hour-a-day basis, to ensure that all data came down in a usable form, and to update the requirements process.35

  In addition to pursuing the modernization program, Huffstutler launched what he dubbed the National Exploitation Initiative, which involved the creation of a National Exploitation Laboratory (NEL). The concept was to create a “users program office” to speak with a unified voice with the procurement program office—the NRO. An early part of the initiative was to invite the directors of the military service’s imagery interpretation units to NPIC, update them on the modernization program, and try to establish a dialogue. The NEL, in addition to serving as a single voice for users in evaluating proposals from the Air Force and CIA elements of the NRO, served as a center for the development of equipment to be used in the exploitation process—helping create the best monitors, software, and other components of modernization.36

  Huffstutler also initiated a database audit to determine the adequacy of the national imagery files, which had never been graded. The NPIC director arranged with the other imagery centers—those in DIA and the military services—to go back and sample past imagery and evaluate the accuracy of the reporting in the written cables. The Strategic Air Command assumed responsibility for the process with respect to air defense imagery, and NPIC took the leadership role for the remainder of the imagery. The audit revealed 96 percent accuracy. Plugging some of the holes discovered in reporting raised accuracy another 1 percent two years later, according to NPIC’s customers.37

  Beyond the problems of modernization, Huffstutler’s NPIC had one major world crisis to deal with—one that required around-the-clock operations.38 On Saturday, April 26, 1986, a nuclear accident occurred eighty miles from Kiev, at the Chernobyl nuclear power plant’s reactor No. 4. As the result of a series of safety violations—including running the reactor without the emergency cooling system and removing too many control rods—a small part of the reactor went “prompt critical.” The effect was the equivalent of a half ton of TNT exploding on the core. Four seconds later, a second explosion blasted the 1,000-ton lid off the reactor, destroyed part of the building, and brought the 200-ton refueling crane crashing down on the core. A “fireworks” display of glowing particles and fragments escaping from the units followed, setting off thirty fires in the building. In addition, the huge blocks of granite in the reactor core also caught fire, spewing out plumes of highly radioactive fission products.39

  The first solid indication that the United States received concerning the accident was from an official Soviet statement on Monday, April 28. There had been signs of unusual activity around Kiev on Sunday, probably from communications intelligence, but what was happening was unclear. Only the following day was the situation clarified.40

  Once alerted to the disaster, the intelligence community responded by turning its full set of resources on the Kiev area. A VORTEX signals intelligence satellite sucked up all military and relevant civilian communications within several hundred miles of Chernobyl. Due to launch failures in August 1985 and earlier in April, the U.S. space imagery capability consisted of a lone KH-11, 5506. It was reprogrammed to obtain photography of the nuclear reactor at the first opportunity. Its last visit had been almost two weeks before, and the first chance for 5506 to provide imagery came on Monday afternoon. However, given its orbital path, the image had to be obtained from a considerable distance and even with computer enhancement didn’t show much. Even had 5506 been closer to its target, the smoke hovering over the reactor area probably would have obscured the site. The following morning, the distance was still too great to produce a good photo, but by evening the KH-11 had approached close enough to return the first good imagery of the accident site. The picture was reported to be “good and overhead.” Huffstutler recalled the imagery as being “right down the core,” showing the concrete cap blasted right out and helicopters and firemen trying to deal with the consequences of the accident.41

  With the photos in hand, analysts at NPIC began assessing the situation. The photos revealed that the roof of the reactor had been blown off and the walls were pushed out, “like a barn collapsing in a high wind,” said one source. Inside what was left of the building, there was an incandescent mass of graphite. Some tendrils of smoke and the blackened roof of the adjoining building indicated that at some point the fire had been more active. The graphite settled down into a glowing mass, while radioactive material from a pile that had contained 100 tons of uranium was still being vented through the open roof and into the atmosphere.42

  The photos revealed activity in the surrounding areas, activity that was quite rem
arkable given the perilous situation at Chernobyl: A barge was sailing peacefully down the Pripyat River, and men were playing soccer inside the plant fence less than a mile from the burned-out reactor. The photos of the town of Pripyat showed that there had been no evacuation.43

  Among those briefed by NPIC with the satellite photos was the House Permanent Select Committee on Intelligence. “We were shown satellite pictures of the reactor building from before and immediately after the explosion,” committee member George D. Brown Jr. (D.–Calif.) said after a closed-door hearing on Thursday, May 1. “They were dramatic, with the roof beams collapsed and debris scattered around the plant. No bodies were visible,” he added.44

  KH-11 photos taken on Thursday morning May 2 showed no smoke at all, leading an interagency panel to believe that it was possible the fire had been put out. Only the day before, despite Soviet claims that the fire had been extinguished, the panel had predicted that it might burn for weeks. Some analysts believed they detected shimmering over the reactor, suggesting that the graphite was still burning. It appeared that the Soviets were dumping dirt or sand on the fire from helicopters. One KH-11 photo showed a helicopter hovering directly in the plume of the radiation.45

  The satellite photos provided data that enabled NPIC’s analysts to determine the validity of reports that Unit 3, adjacent to the damaged Unit 4, was affected by a meltdown or fire. From further satellite photos, a federal task force concluded on May 3 that the other Chernobyl reactors were not at risk. Lee M. Thomas, head of the task force, announced that on the basis of the photos, “we see no problems with the other units.”46

  NPIC undoubtedly continued to monitor the cleanup, examining KH- 11 images of the Mi-8 helicopters (with lead shields on their floors). The helicopters flew hundreds of missions day after day, dropping sacks of sand through the broken roof of the reactor from heights of more than 650 feet. Workers then sealed the roof shut with tons of lead pellets, which rolled into whatever cracks remained between the sandbags.47

  A VERSUS B

  On January 24, 1985, about fifteen months before the Chernobyl accident, the space shuttle orbiter Discovery was launched on the first secret mission in fifteen shuttle flights, carrying another product of OD&E’s efforts. Earlier that week, Brig. Gen. Richard F. Abel, the Air Force director of public affairs, warned reporters not to speculate about the payload and threatened that the Department of Defense was ready to launch an investigation in the event of any unwelcome stories.48

  But times had changed dramatically since Air Force representatives could get away with telling reporters (as they did for the first CANYON launch) that they wouldn’t want to know the nature of the payload. They certainly did, and the warning did not stop the Washington Post from revealing that the secret payload was a new SIGINT spacecraft. The spacecraft had been developed under the code name MAGNUM, but that name had been changed to ORION by the time of launch. Most important to the U.S. intelligence community was that the shuttle’s military astronauts were able to release the satellite safely into space, so that the rocket attached to it could place it in its proper geosynchronous orbit.49

  The spacecraft, which weighed about 6,000 pounds, was apparently placed in orbit over Borneo. It was reported to have two huge parabolic antennae, one for intercepting communications and telemetry signals and the other for relaying the signals to the ground station at Pine Gap, Australia. ORION was the successor to the RHYOLITE/AQUACADE program, and an improvement both on RHYOLITE and ARGUS, its intended successor. Compared with RHYOLITE, it could pick up lower-powered signals, such as “turned-down” telemetry signals, intercept a wider range of frequencies, and due to its bigger transmitting antenna focus its signal more sharply to its ground station. In addition to intercepting signals never intended for U.S. intelligence analysts, it apparently also received data from emplaced sensors in the Soviet Union and elsewhere.50

  In 1985, not long after celebrating the successful launch of ORION, Robert Kohler, the former KENNAN program head who in March 1982 succeeded Bernie Lubarsky as chief of the Office of Development and Engineering (i.e., NRO Program B), found himself in a battle with the Air Force and the NRO over the plans for the next SIGINT satellite program.

  A native of Rochester, New York, Kohler had considered the snowiest city in the United States no place to live. He graduated from the Rochester Institute of Technology, married, and left town in about a week, refusing to interview at the Rochester-based Xerox and Kodak corporations. His first exposure to space reconnaissance was as head of the photo science technology department at Itek in Boston, where he worked on the CORONA program. In 1967, he joined the CIA for what he expected would be a few years of government service before returning to Itek. When time came to return, he decided to stay at the agency because he was “having too much fun.”51

  Kohler was certainly not one to avoid a battle when he believed someone else was headed down the wrong path. The Air Force Office of Special Projects (NRO Program A) and its director, Ralph Jacobson, was proposing that the follow-on to ORION be a far bigger system, which they believed was necessary to cope with a new type of Soviet microwave signal.52

  Kohler thought it was feasible to take the system then in operation and modify each successive satellite, never spending too much money at any one time. He also believed that NRO director Pete Aldridge felt that with a budget squeeze on the horizon, it might be his last chance to get approval for a major new initiative. In addition, according to Kohler, the CIA’s Program B had continually bested the Air Force’s Program A during Aldridge’s tenure, and Aldridge wanted to give Program A a victory.53

  Kohler’s position as head of development and engineering was somewhat different from those of his predecessors in earlier eras. When Kohler took the job, John McMahon, then deputy director of central intelligence, told him he should act as if he worked for the director of the NRO—unless he believed that DCI William Casey was not being adequately served by the NRO director. In 1985, Kohler felt that caveat applied—particularly when Aldridge did not present Kohler’s alternative to Casey.54

  Made aware of the differences, Casey asked to hear from both Air Force and CIA elements of the NRO. Two critical meetings followed involving Casey and Secretary of Defense Caspar Weinberger. McMahon asked to see the briefing Aldridge was planning to give and told the NRO director that it should not contain a recommendation. But at the first meeting, Weinberger asked Aldridge for his recommendation and got it. Casey demurred making a decision then, telling Weinberger that he had not had an opportunity to study the options. When Casey was briefed by the NRO Staff, he found they omitted the Program B option. Again he refused to make a decision.55

  The following Monday morning, Kohler drafted and sent a four-page memo to Casey through Hineman. According to Kohler, he faced opposition not only from Aldridge but also from the CIA’s Executive Director, former DS&T associate deputy director James Taylor, who wanted Casey to accept Aldridge’s option. The memo laid out the NRO briefing as well as what was omitted and invited a decision. Instead of a decision, Kohler got a call that night from Casey, telling him he wanted to see him the next day.56

  Casey, according to Kohler, “had ignored the DS&T for the first four and a half years of his tenure.” The next day, he asked the development and engineering chief to tell him how signal recognition worked. After ten minutes, Casey walked over and opened the door to McMahon’s adjacent office and asked his deputy to join them. They talked for several hours, with Kohler addressing the issue of the intelligence value per dollar from each of the proposed SIGINT systems, using different scenarios constructed by the intelligence directorate.57

  The presentation convinced Casey to consider the Program B proposal. He transmitted a sanitized version of Kohler’s memo to Aldridge, PFIAB chairman Anne Armstrong, and two technical people on the board—John Foster and Bud Wheelon. Kohler then heard from McMahon, who told him to expect a call from NRO deputy director Jimmie Hill, who would arrange for him to brief Anne Armst
rong. When twenty-four hours went by without the call, Kohler called McMahon, who then sent him to La Jolla, California, to see the PFIAB chairman.58

  Wheelon later recalled first speaking to Casey on his way from Denver to Aspen to attend an Aspen Institute summer study meeting, apparently sometime after Kohler left the agency in August 1985. He considered it “an unwelcome assignment,” which couldn’t do the Hughes company any good, but he agreed to meet with Aldridge and some of his staff at a motel near the Aspen airport. Wheelon examined the issue, relying on Hughes official Harold Rosen as technical backup.59

  Wheelon concluded that extensive improvements suggested by Program A were not necessary—that its staff had missed two key technical points—and that the CIA proposal to stay with the same basic system and same contractor (TRW) made the most sense. He spoke to Jacobson about the matter, and the Program A head accepted his analysis.60

  Two years later, it seemed Kohler’s effort had been wasted. Casey had approved the Program B alternative, but he died in 1987. Robert Gates became the acting DCI, and although he had supported the program as chairman of the Intelligence Producers Council, he decided to sacrifice the program to budgetary requirements. But the program’s death would be only temporary.61

  KODIAK

  In 1985, Kohler was succeeded by Julian Caballero, who would remain as office director until fall 1993.62 A system on the drawing board when he assumed command was one dedicated to controlling and relaying data for the entire reconnaissance fleet. The proposed system, originally code-named IRIDIUM and then KODIAK, would have consisted of four geosynchronous satellites, with one satellite in view of the Washington-area downlinks at Ft. Belvoir and Ft. Meade. That satellite would have been capable of downlinking the information in such a narrow beam as to make it virtually immune to interception. The other three satellites, in addition to their control functions, would be able to receive data from reconnaissance satellites and then transmit the data to the downlink satellite or, via a laser crosslink, to another satellite that would then transmit the information to the downlink satellite. Because of funding limitations, the proposal was killed in 1987. Funding for KODIAK would have required cuts in the Strategic Defense Initiative budget that the administration was unwilling to make.63