Eye in the Sky: The Story of the CORONA Spy Satellites

by Dwayne Day

  To staff the new facility, the USGS management gave us the freedom to visit our field offices and conduct personal interviews with employees to determine their interest in the CORONA effort. One can imagine the potpourri of individuals we amassed. Some people were innovators, others were individuals who wanted to learn, and still others wanted to use this unique experience as an opportunity to advance in the scientific community. Well, we brought these people together and managed to begin revising USGS maps in about three or four months after opening the new classified facility.

  The first CORONA materials we received were KH-3 images, which we desperately needed in order to revise the 1:250,000-scale maps of the United States. These maps, which provided full coverage of the continental United States and Alaska, were originally prepared by the Army Map Service during World War II and ultimately completed about 1950. The Aeronautical Chart and Information Center (ACIC) in St. Louis provided the procedures for compiling the maps and rectifing the CORONA materials. Bill Mahoney deserves much of the credit for the innovative developments that occurred in the technology we were working with at the time.

  The National Photographic Interpretation Center (NPIC) was a great help to the USGS. They opened their doors to the USGS in many ways. They provided training, equipment and great insight into the processes required to interpret space imagery. NPIC’s director, the late Art Lundahl, specifically got involved and provided surplus equipment to the USGS for mapping applications. Thus, the total USGS operation was built on a shoestring—with the cooperation of many governmental organizations.

  The advantage USGS had over the military and the other segments of the Intelligence Community was that we had excellent “ground truth.” The original maps were accurately compiled and fitted to a well-developed project and geoid. Thus, we obtained the best rectified imagery available to us, used “green-eyeshade” cartographic processes, and revised the 1:250,000-scale series for the contiguous U.S. and a great portion of Alaska. Some images of Alaska, however, were of poor quality because of the orbital inclination of the spacecraft and the cloudy conditions that often exist in that part of the world.

  In 1970/1971, another milestone occurred that stimulated the use of CORONA materials in civil programs. The Bureau of the Budget formed the Federal Mapping Task Force, which was commissioned to take a look at all mapping activities in the United States. The creation of the task force encouraged some other civilian agencies to get more deeply involved in the peaceful exploitation of CORONA materials.

  Some competition consequently occurred among various groups.14 The USGS was able to obtain modest line-item funding, which in turn accelerated the capital equipment program and helped the USGS obtain state-of-the-art hardware and software. There was also an unofficial group formed in 1971 called the Civil Applications Committee (CAC), which interfaced with COMIREX. The CAC was able to take the engineering test flights, the principal source of information for civilian organizations, and closely coordinate and prioritize them. The CAC was chaired by the USGS, and served to coordinate civilian priorities until the end of the CORONA era. It continues to function today.

  As technology matured, we realized another important mapping objective could be satisfied by using CORONA materials. We concluded that it would be possible to inspect how up-to-date were the 55,000 quadrangle maps of the continental United States at 1:24,000 scale (1 inch equals 2,000 feet). The process was very simple. We projected the CORONA images over the maps to ascertain whether or not they would need revision, based on the extent of the changes shown on the satellite image. For example, a trained employee could view a roll of CORONA film and quickly annotate the related maps by using USGS criteria to decide whether or not the maps would require revision. This process was so reliable that we later just reprinted some of the inspected maps and included the statement “No Revision Required, Inspected 19__.” Of course, maps which required revision would be corrected by conventional methods.

  In summary, there were three major accomplishments that were realized as a result of the interaction with the CORONA program. First, the 1:250,000-scale maps of the continental U.S., and parts of Alaska, were revised once, and in some cases twice. Second, several thousand 1:24,000-scale quadrangle maps were inspected by using more than 40,000 CORONA photographs. And third, and most important, we formed a cooperative relationship with the Intelligence Community and with other government agencies, which probably would not have occurred without the CORONA program.

  As Gifford, Mahoney, Daugherty, and Starr attest, CORONA had a revolutionary impact on the military and civilian mapmaking processes at a key point during the Cold War. Due to the cartographic innovations that geodesists and photogrammetrists developed, and the tremendous cooperation between the civilian, military, and industrial groups involved in the CORONA program, mapmakers were able to create a sophisticated world geodetic coordinate system that has had a profound effect on the maps that defense planners and civilians have continued to use.

  However, perhaps Gifford and Mahoney’s final perspectives are the best way to summarize the revolutionary influence CORONA had on mapmaking, for, as Gifford notes:

  Tremendous things occurred during the CORONA program. There were wonderful people doing great things in the mapping community despite the odds. After all, we faced a very demanding customer; the Defense planning community wanted the information yesterday. So photogrammetrists and geodesists had to deal with the continual stress of providing better and more accurate information in a timely fashion. We were all racing the clock. But we never had a sense that we were doing it in isolation; there was always high priority put on all of the activities that supported the mapping and dimensional areas of the CORONA system.

  And that prioritization apparently paid off, because as William Mahoney notes about the mapmakers who worked on the CORONA program:

  We provided the last essential element. We put the cross hairs on the target that made our counterforce credible.

  11

  EXPLOITING CORONA IMAGERY

  The Impact on Intelligence

  American imagery and intelligence analysts’ work changed drastically as a result of the CORONA program. In this chapter, four former analysts discuss the significant impact that the CORONA project had on U.S. intelligence and comment on a wide range of issues, including how CORONA began and its influence on American strategic planning. The heart of their discussion focuses on the techniques and strategies that they developed to exploit CORONA imagery so that they could provide government officials with up-to-date and accurate information. They also highlight the actual discoveries they made and challenges they faced. In all, the four analysts conclude that the CORONA program was an absolutely essential component in the successful gathering and assessment of intelligence during the Cold War era.

  Robert (“Rae”) M. Huffstutler, a former intelligence analyst and manager at the CIA’s Directorate of Intelligence during the CORONA project, provides an explanation of why the Intelligence Community desperately needed a satellite reconnaissance system like CORONA in the early 1960s.1 He also outlines some of the questions policymakers wanted answered once CORONA became operational. For example, Huffstutler notes:

  Starting in 1949 and going through the 1950s, the Soviets, in their denied area, had a large series of successful nuclear tests. There were long-range bombers rolling off the assembly lines in quite uncertain numbers. There were ballistic missile tests, but we did not know whether there were actually ballistic missile sites, though there were a lot of reports that there were. And the 1961 launch of Yuri Gagarin and his space ride simply heightened the anxiety held by national security policymakers and caused them to want the answers to some fairly straightforward questions, which were a lot easier to ask than they were to answer. The initial questions that they were asked were of a “yes” or “no” variety. Is there a missile gap? Is there a bomber gap, and, if so, to what extent does it exist?2 For the answer to these questions, we had to rely very heavily on remo
te sensing, because it was our only look into a denied area—a police state where travel was carefully controlled and constrained.

  Before CORONA, American intelligence about the Soviet Union and other countries was very fragmented. Richard J. Kerr, a CIA analyst who worked at the Office of Current Intelligence during CORONA’s early years, explains just how difficult it was for intelligence officers to make accurate assessments before the advent of CORONA imagery. He also hints at how their analytical methods changed once they started getting such information.3 Kerr remembers those years as

  an extraordinarily challenging period. You know for a good deal of time we didn’t know what missile sites looked like. And we didn’t know their pattern. And we didn’t know what a gaseous diffusion plant or a reactor looked like. Because at the beginning of this process, it was kind of a clean slate. We knew what we knew about ourselves and we had all this information—some of it out of the SIGINT [signals intelligence] world—and we had a lot of information from refugees about plants and facilities on little 3 by 5 cards.

  So it started up like a puzzle, but we didn’t have the big picture of what the puzzle was. So we’d put a piece down and then we’d try to figure out whether it was the corner, the bottom, the top, and then we’d try to figure out where it fit in. And for several years, we were in that business of trying to fit together these pieces, and a lot of people put those little pieces together and developed the techniques to put them together. What the people involved were doing was trying to make sense out of a huge variety of information and photos, trying to connect those installations and connect the patterns. They were trying to make sense out of it.

  We had some idea of what the rough structure was, but to my knowledge we had not accurately identified the location of any major ICBM facility prior to the availability of CORONA imagery. And, in fact, some of the things that we identified as “highly probable” turned out not to be missile sites. I think Novorosissk was one of the areas that we zeroed in on and said this has got to be one. Well, it wasn’t.

  As this comment suggests, once CORONA began successful operations in August 1960, it started providing analysts with “new information,” which, as Rae Huffstutler recalls, allowed them to sharply revise the National Intelligence Estimate of the Soviet Union’s ICBM capabilities, and to carry out more strategic planning. Huffstutler notes:

  Just to give you an idea of the impact of imagery on the intelligence estimating process, I pulled three quotes from three National Intelligence Estimates [NIE] which have recently been declassified. These are all from an estimate designated “11-8” which was a series on “Soviet Capabilities for Strategic Attack.” The first one was published in February of 1960, and in its conclusion it said that the Soviet ICBM program would probably provide on the order of 140 to 200 ICBMs on launchers in mid-1961. The basis of this estimate was the reporting of some immigrants, what HUMINT we could collect, and SIGINT. But this information was unclear about the total numbers and disposition of the forces.

  The Yurya ICBM complex, USSR, June 28, 1962. Soviet ICBM facilities were a primary target for CORONA missions. They were easily discernible by the multiple security fences that surrounded them. (Photo courtesy National Photographic Interpretation Center)

  In August 1960, some six months later, a conclusion was published in the next NIE 11-8, which laid out three illustrative programs using different assumptions. By this time nobody had any confidence in the facts and they tried to show that if the Soviets produced as fast as we thought they could produce, they could have 50 to 200 ICBMs on launchers by mid-1961.

  However, by September 1961, there was a major change in NIE 11-8. It concluded that new information had caused a sharp downward revision in the estimate of Soviet ICBMs on launchers to between 10 and 25. In fact, I think in retrospect, we found out there were only six operational sites, plus some launchers at the test range. Obviously the difference between the patterns of evidence that you get with and without imagery is significant.

  On the medium-range and the intermediate-range missiles, the fact is that we had a lot of accurate information because—being of shorter range—they were in the western

  Soviet Union, and we had better access to immigrants and to reporting coming out of that area. But the impact of imagery alone makes a difference—a difference you can see when the Intelligence Community has factual data instead of speculative theories. For example, with the acquisition of hard evidence, the size of the National Intelligence Estimate went from an inch thick down to about only ten or fifteen pages. That was a particularly dramatic time in the analytical business because we could tell national security policymakers basically what the strategic balance was. In short, how bad the threat that we faced was. For the disposition, location, and size of strategic forces, there was simply no substitute for the CORONA system.

  Well, where did all this lead? Once you could describe not only the strategic threat to the United States, but also the ground threat to the United States, and once you could say where the threat was positioned, you could begin to describe how large a mobilization would be necessary for the Warsaw Pact to overrun Europe. This was the basis for strategic warning and allowed senior policymakers in Washington to do strategic defense planning. There was no need to hedge against imaginary forces when we knew what we were dealing with on a factual basis, especially when we knew the size and disposition of all operational military formations.

  Eventually this also allowed us to pursue strategic arms limitations. After all, it would have been impossible to establish sufficient political confidence in a verification process based only on HUMINT and SIGINT sources, especially when you always had to assess whether a HUMINT source really knew what he or she purported to know. You had to ask yourself if he or she had the whole story.

  During CORONA’s early years, image analysts developed a number of techniques and methods to exploit CORONA’s imagery to its fullest extent. In the following section, Dino Brugioni, a senior manager at the National Photographic Interpretation Center (NPIC) during the CORONA program, details the interpretation process that analysts created and the problems they had to overcome during the program’s tenure.4

  The image analysis process is painstaking and often tedious. With the advent of multi-sensor technology it has grown both in complexity and sophistication. It is largely cognitive and is based on the recognition of features and patterns that are of special interest to intelligence. These special features are called signatures. For example, SA-2 surface-to-air missile sites display a Star of David pattern. An SS-5 intermediate-range ballistic missile site, in the initial stages of construction, is often referred to as “slash marks.” All of these signatures are carefully cataloged and often referred to as “P.I. [photo-interpreter] keys.” The imagery analyst develops a unique combination of skills and knowledge. Inherent traits include attention to detail, curiosity, inquisitiveness, diligence, deductive and inductive reasoning, and above all, good eyesight.

  The use of radar, infrared, and multispectral sensors during the Vietnam conflict required interpreters to become sensitized to new forms and patterns. An experienced analyst could derive an enormous amount of information from what might have seemed to be an out-of-focus, improperly developed photograph.

  Arthur Lundahl, first director of the National Photographic Interpretation Center (NPIC). Lundahl built NPIC into a major center of expertise on the interpretation of overhead photography. (Photo courtesy Dino Brugioni)

  Over the years, the experienced analyst has cataloged hundreds of methods to outfox the enemy. He knows that man lives by laws, rules, customs, and practices. [Arthur] Lundahl [director of NPIC during the CORONA era] frequently addressed his new interpreters with this analogy: “Scribe a 25-mile circle on a map in most areas of the world, and man is born, lives and dies within that circle. Carefully analyze the aerial photographs of that circle, and you’ll determine what man eats, what he wears, his source of water, what he cooks with, where and how he’s educ
ated, his customs, how he makes a living, his religion, the home he lives in, his interaction with nature, and finally, where he’s buried.”

  In exploiting imagery throughout the year, we became acutely aware that each day, week, month, and season presented serendipitous benefits that could be exploited. For example, the best military order of battle is obtained on Sunday morning when most of the equipment is in garrison. Capabilities of ground forces are best observed in the spring training exercises. A heavy snowfall negates all camouflaging efforts. By tracking snow-clearing operations of military facilities, one gets the idea of the importance of each building. The headquarters building is usually cleared first, followed closely by clearing the paths to the latrines. Melting snow on a roof indicates which buildings are heated and activities in the forest are best observed in the spring before the trees begin to leaf.

  Security fences which prevent ground observation of an installation are also an immediate flag for the imagery analyst. The Soviet penchant for security immediately aided our interpretation effort. Soviet strategic installations were ringed by not one, but by several fences and these were very apparent.5

  Although atmospheric effects frequently degraded CORONA imagery, they could occasionally make the photo-interpreter’s job easier, as in this photograph, where a surface-to-air missile site is revealed by snow cover. This site was imaged by CORONA Mission 1029, launched on February 2, 1966.

  When the Soviets began shipping arms and munitions to other countries, and they knew we were overflying them, they built crates to transport and conceal the military hardware they were sending abroad. Our photogrammetrists could precisely measure the crates and determine what was in them. So we developed “crateology” as part of our function. We also developed “tentology,” so we could count the number of tents and get an idea of the number of troops in an area. There were all kinds of these little techniques that were developed over this period.