The Wizards of Langley
Page 5
Several years before the campaign, by the time the National Intelligence Estimate (NIE) “Main Trends in Soviet Capabilities and Policies 1957–1962” was completed in November 1957, the Soviets had already tested an ICBM, and it was feared they might have about ten prototype ICBMs for use by 1959. Then, in a December 1958 estimate, the intelligence community stated its belief that the Soviets intended to acquire a sizable ICBM force at the earliest practicable date. The NIE also pointed out the absence of sufficient evidence to judge conclusively the magnitude and pace of the Soviet ICBM program. Indirect evidence, including production capacity and the ability to construct launch facilities, to establish logistic lines, and to train operational units, led to the conclusion that the Soviets “could achieve an operational capability with 500 ICBMs about three years after the first operational date [1959].”122
Over the next several years, that judgment was revised downward, in the absence of intelligence to sustain earlier high estimates of the pace of Soviet ICBM deployments. The 1959 NIE suggested that the Soviets might have 140 to 200 ICBMs on launchers by mid-1961, and, speculatively, 250 to 350 by mid-1962 and 350 to 450 by mid-1963. Though smaller than previous estimates, they were consistent with a Soviet missile force that could destroy the vulnerable strategic bomber bases of the Strategic Air Command, particularly since it was believed that improvements in the accuracy and reliability of Soviet ICBMs had sharply reduced the number required to launch an effective attack.123
Although such estimates were highly classified, their basic thrust reached the public through key columnists such as Joseph Alsop and prominent politicians such as Senator Stuart Symington (D.Missouri). The electorate was well aware of Soviet missile tests and even more so of the Sputnik and other Soviet space launches. The expectation of a substantial missile gap in the Soviets’ favor thus became part of the 1960 election, with Democratic candidate John F. Kennedy lambasting his opponent, Richard Nixon, over the issue.124
The NIE for Soviet strategic offensive forces during the period 19601965, issued the month after Kennedy’s razor-thin victory, reflected different views by various intelligence organizations, but was consistent with the expectation of a significant Soviet advantage in the near future. The Air Force took the most pessimistic view, predicting 200 ICBMs by mid-1961, some 450 by mid-1962, and 700 by mid-1963; the CIA predicted 150, 270, and 400 for the same periods. At the low end were the Air Force’s military rivals. The Army and Navy jointly predicted deployments of 50, 125, and 200.125
In June 1961, the intelligence community issued a new assessment. It argued that the Soviets might have 50 to 100 operational ICBM launchers and therefore the ability to bring all SAC operational air bases under attack. In any case, the estimate concluded that the Soviets would have 100 to 200 operational launchers within the next year and would almost certainly be able to attack then.126
By September 1961, new intelligence had a dramatic impact on the estimates. One item was a top-secret report, “The Soviet ICBM Program,” that was based on information from an officer in the Soviet military intelligence service, Lt. Col. Oleg Penkovskiy; this document sharply discounted the near-term missile threat. In addition, electronic monitoring of Soviet missile and space test centers provided data on the types of missiles being developed. The third, and most conclusive, source was CORONA photography.127
The estimators noted that through CORONA operations since mid1960, “our coverage of suspected deployment areas in the USSR has been substantially augmented” and that the photography had “been studied in detail by photo-interpreters with knowledge of US and Soviet missile programs.” Analysis revealed that many of the suspected areas did not contain ICBM complexes as of summer 1961. Thus, the NIE, which drew a predictable Air Force dissent, estimated “that the present Soviet ICBM strength is in the range of 10–25 launchers from which missiles can be fired against the US, and that this force level will not increase markedly during the months immediately ahead.” The expected number of Soviet ICBMs by mid-1963 was 75 to 125. The Soviets had apparently chosen to deploy only a small number of heavy and cumbersome first-generation SS-6 ICBMs and to concentrate their efforts on a smaller second-generation system for deployment, probably in 1962.128
The people who had been members of the CIA’s Photographic Intelligence Center when the CORONA program yielded its first imagery were, by September 1961, employees of the National Photographic Interpretation Center (NPIC). NPIC was formally established by National Security Council Intelligence Directive (NSCID, pronounced N-Skid) No. 8 of January 18, 1961.129 NPIC was to be run by the CIA as a service of common concern for the entire intelligence community—interpreting both aerial and satellite imagery. Its establishment provides another example of how far the CIA had come in employing science and technology in the pursuit of intelligence requirements.
NPIC replaced the PIC and was put under the command of Arthur Lundahl, who already had a long career in the photointerpretation business. At the University of Chicago, he majored in geology. His involvement in aerial photography began during his years as a research assistant, performing geological research in Ontario, Canada. His work required the production of maps based on aerial photography.130
Because of his background, Lundahl was requested, as war approached, to help train photointerpreters. After conducting training courses and obtaining his M.S., he joined the Navy and worked on photointerpretation, both in Washington and overseas.131
At the end of World War II, he found himself back in Washington, and in 1946 he was asked to help write the charter for the Naval Photographic Interpretation Center. After seven years at the Center, he became convinced that the Navy was “going to go nowhere in photointerpretation as it was then structured.” He was also approached by the CIA, which had become aware of a paper he had written—“Consider the Mata Hari with Glass Eyes”—that focused on the potential role of photographic intelligence during the Cold War. His initial reaction was one of skepticism. Years later he recalled telling CIA official Otto Guthe, “I don’t know anything about you guys. If you’re going to parachute me into Salerno or somewhere, forget it. I’m a scientist.” Guthe reassured him that the CIA “was going into the photointerpretation business and they wanted someone to come over and run it who had experience and the right credentials for doing the job. They wanted me to do it.”132
As a result, Lundahl joined the CIA in early 1953, as head of the Photographic Intelligence Division (PID) of the Office of Research and Reports—with thirteen people and a few hundred square feet of floor space over a Ford dealership, in the Steuart Building, at 5th and K Street, N.W. The operation had the CIA code name HT/AUTOMAT—HT from the initials of security officer Henry Thomas and AUTOMAT because Lun-dahl conceived of the division as a place where intelligence consumers could come and pick up whatever they needed in terms of interpreted photography. Lundahl recalls that “there was a very narrow front. It was very inconspicuous. It had . . . some funny little sign on the door, and there was some kind of turnstile about 100 feet in from the door and a glass cage where a security guard sat and a dirty little elevator which ran slowly at best. Of course, there was no place to park and no place to eat.”133
Expansion came a few years later in the wake of the U-2 project, and by 1956, 150 interpreters were occupying 50,000 square feet in the Steuart Building. In August 1958, the PID was merged with the Statistical Branch of the Office of Current Reference and became the Photographic Interpretation Center.134
With the initial expansion to 150 personnel came military people detailed from the various armed services. Lundahl sought to make them feel like coproprietors of a national facility, since he felt that if he “had Army, Navy, Air Force, State, NSA, CIA people in there, I would have all the ingredients for nationalization, although this was nowhere in the cards in 1956.”135
Several years later it was in the cards. In March 1960, defense chief Thomas Gates suggested to Eisenhower the need for a study of the defense intelligence establishmen
t, which he described as a huge conglomerate spending $1.52 billion annually. Although Eisenhower reacted favorably to such proposals, the necessary impetus was the shooting down of the U-2 in May.136
A May 6 meeting between Allen Dulles, Gates, Budget Bureau chief Maurice Stans, presidential adviser Gordon Gray, and the President’s Board of Consultants on Foreign Intelligence Activities resulted in the creation of the Joint Study Group to review various aspects of the U.S. foreign intelligence effort.137 Recommendations contained in the group’s December 15, 1960, report covered all aspects of U.S. intelligence operations—collection, analysis, and the roles of the DCI and the United States Intelligence Board (USIB). Several recommendations were based on the many hours the group spent discussing the problem of processing and interpreting overhead photography. The report noted agreement “in most of the community that a central photographic intelligence center should be established,” although it noted that opinions varied as to how much interpretation and analysis should take place at such a center and who should run it. The report recommended that the DCI and Secretary of Defense should determine the details concerning the center’s management and that a National Security Council Intelligence Directive should be drafted establishing a National Photographic Interpretation Center.138
The question of who would run the center was debated, but not resolved, at three USIB meetings. As a result, the issue was brought up at the January 12, 1961, meeting of the National Security Council. Participants included Eisenhower, Gates, Joint Chiefs of Staff (JCS) Chairman Lyman Leminitzer, Allen Dulles, and presidential science adviser George Kistiakowsky.139
The center would be run by either the Department of Defense or the CIA. Gates and Leminitzer argued for Defense because of the military’s role in the collection of imagery as well as being the primary intelligence consumer. In addition, Gates argued that although Dulles agreed that Defense should take the lead in the event of war, continuity was required for a proper transition from peacetime to wartime operation. He gave assurances that a Defense-run center would not be removed from Washington and would provide its services to the agencies outside of Defense that required them.140
Dulles countered that the PIC was a joint enterprise with 100 Army, 10 Navy, and 7–15 Air Force officers. In addition, the information produced by the center was chiefly military only with regard to targeting, whereas photographic intelligence had tremendous political significance and was a matter of common concern to a number of Washington agencies, including the Department of State. Furthermore, the center had developed a group of career officials who intended to make a career of photographic intelligence. In contrast, military officers were regularly rotated to new jobs. When Gates suggested that if the center was placed under Defense a career staff would be retained, Dulles responded that abandonment of rotation was a new idea for the military.141
When Eisenhower expressed his concern that the military had a greater need for timely information than the civilian agencies, Kistiakowsy told him that the CIA center was providing the military with the material it required without delay, that the existing center was “a revolution in photographic techniques,” and that the CIA had taken the lead in managing and developing the center. To disturb the arrangement would result in delay and a loss of progress. Instead he suggested expansion of the center.142
At least partly on the basis of Kistiakowsy’s advice, and partly because he believed Defense had not shown any unhappiness with the current arrangement, Eisenhower decided that the CIA should continue running the interpretation center and that it should be expanded. According to Lundahl, Eisenhower’s words were simple: “Well that settles it, Allen . . . Allen Dulles, you’re going to run this thing, so carry on.”143
QUALITY ELINT AND CLANDESTINE COMINT
In late May 1954, while the CIA was in the final stages of preparing for a major covert-action operation designed to remove the leftist president of Guatemala from power, Allen Dulles briefly turned his attention to a very different aspect of the agency’s activities. On the twenty-ninth, he approved the CIA’s first electronic intelligence (ELINT) program.144
The potential value of intercepting and analyzing the signals from radars under development and missiles during flight testing had been recognized by several officers in OSI—particularly James Spears, Ralph Clark, and George Miller. In 1951, OSI had begun building up its expertise in the field as well as giving technical assistance to the British ELINT program. The NSC provided an additional stimulus in 1953, when it assigned the CIA responsibility for evaluating Soviet capabilities for jamming radio signals.145
At the time, OSI, which analyzed the material collected, was not the only agency component involved in ELINT. The Deputy Directorate for Plans (DDP) obtained ELINT data through liaison relationships with foreign intelligence services as well as through its own clandestine collection operations. The Office of Communications, a component of the Directorate of Administration, designed and produced equipment for the collection and processing of electronic signals. In October 1953, to evaluate the CIA’s activities in the area, an ELINT Task Force was established to review the agency’s efforts, which resulted in the plan approved by Dulles. Among its provisions was creation of a CIA ELINT Staff Officer (ESO) to coordinate the agency’s ELINT activities, a post held for many years by a member of OSI.146
The roles of DDP, OSI, and the Office of Communications remained essentially unchanged throughout the 1950s. But by 1959, there was a substantial expansion of the CIA’s ELINT operations, which was facilitated by the National Security Agency’s relative lack of interest in the area. NSA had been established in 1952 to perform the national signals intelligence (SIGINT) mission, with power to coordinate the SIGINT activities of the military services. But it was most interested in communications intelligence—obtained by intercepting voice and fax traffic. CIA officials who required ELINT believed NSA slighted the activity in allocation of assets and personnel. In particular, they felt that NSA paid insufficient attention to the collection and analysis of Soviet missile telemetry—which was crucial to estimating Soviet missile developments.147
Collecting such signals would require cooperation with countries in close proximity to Soviet testing facilities. On December 8, 1954, a U.S. delegation, led by the CIA’s Frank Rowlett, began formal negotiations for a U.S.-Norwegian communications intelligence agreement. Two days later the negotiations concluded successfully. Cooperation involved both COMINT and ELINT, with the CIA funding some of the Norwegian operations. In summer 1955, the Norwegian Defense Intelligence Staff (FO/E) began collecting ELINT from a site at Korpfjell, near Kirkenes and about two miles from the Soviet border. Converting the site to permanent status began in August 1957, and the resulting station, codenamed METRO by the Norwegians, began operations in 1958.148
METRO’s targets included some of the telemetry from Soviet naval missile tests in the far north. In 1958, the Soviets began deploying the SS-N-4 Sark missile on its Golf I submarines. The short-range (350 nautical miles) ballistic missiles carried a single warhead with a yield of 2–3.5 megatons. In 1960, Korpfjell became the responsibility of the newly created Defense Experimental Station Kirkenes, which received its funding from the CIA.149
ELINT cooperation with the Norwegians was not restricted to land. In June 1955, the Norwegian sealer Godoynes began a 69-day mission in the Barents Sea, code-named SUNSHINE, with the objective of gathering signals from Soviet radars as well as intercepting communications. According to Ernst Jacobsen of the Norwegian Defense Research Establishment, who designed some of the Norwegian monitoring equipment carried by the ship, the vessel was also “bursting at the seams with modern American searching equipment, operated by American specialists.” The modern American equipment came from the CIA.150
In 1956, the Norwegian intelligence staff concluded an agreement with the shipping company Egerfangst Ltd., whose ships were then employed to collect communications and electronics intelligence in the Barents Sea and along the coastline of
the Kola Peninsula. The following year, the former fishing boat Eger became the first custom-built ELINT ship. Its transformation was completely paid for by the United States, with the CIA supplying its intercept equipment. The Eger then spent several months in the Barents Sea, gathering SIGINT on Soviet naval activities on behalf of the CIA.151
Also in the late 1950s, the CIA established a telemetry intercept station in Iran on the southeastern corner of the Caspian Sea. Intelligence analysts within the CIA as well as outside advisers, such as future Secretary of Defense William Perry, then at Sylvania, wondered if the low-powered signals that the Soviet missiles would transmit twenty miles to stations on Soviet territory could also be intercepted 1,000 miles away in Iran. CIA officials gained the Shah’s permission for a contingent of technicians to set up antennae in an ancient hunting castle at Beshahr to conduct hearability tests. The tests were to determine how well signals from the Tyuratam test range, which the CIA believed would become a major Soviet test facility, could be intercepted.152
In 1958, as part of the “Quality ELINT” program, the CIA made the first significant attempt to measure the power of a radar for intelligence purposes. The targeted radar was the Bar Lock, which had been rapidly and extensively deployed in East Germany and other areas on the Soviet periphery. Intelligence indicated that the new radars were being used to detect and track U-2s penetrating Soviet territory.153
Photographs of the Bar Lock indicated that its power output might be sufficiently great to significantly improve Soviet aircraft detection and tracking capabilities. Given the potential threat to U-2 operations, those managing the program pressed for more conclusive intelligence on the Bar Lock’s capabilities. Equipment capable of measuring the power output of a radar was installed in a C-119 aircraft. A series of flights followed, the plane flying ostensible supply missions through the air corridors to Berlin, where Bar Lock signals were easily intercepted. Although not all the measurements taken were of high accuracy, the accumulated data did indicate that the power output of Bar Lock was far less than had been determined from photos of it, a conclusion later confirmed by other sources.154