The Bird’s-Eye View

It is obvious that in observing the weather through the “eye” of a high-altitude robot almost all the quantitative measurements usually associated with meteorology must fall by the wayside.

—From a Project RAND report: Inquiry into the feasibility of weather reconnaissance from a satellite (1951), by William Kellogg and Stanley Greenfield.

D

uring World War II, Japanese paper balloons floating on currents in the upper reaches of the lower atmosphere carried incendiary bombs across the Pacific to the United States. They caused some forest fires, which were quickly extinguished. Censorship kept news of the few fires from the public, and thus the balloons did not precipitate the widespread consternation that Japanese strategists had hoped for.

William Kellogg and Stanley Greenfield were intrigued by the story of these balloons and were impressed by the knowledge of the atmosphere that such a campaign had needed. The two men worked for a newly formed group of technical consultants known as the RAND Corpora­tion.[10] The interest the two men had in high-altitude balloons, which they believed might make good platforms for photo reconnaissance and intelli­gence gathering, evolved eventually into a conviction that satellites would provide a good platform from which to observe the weather. Their early conceptual work on “weather reconnaissance” became part of TIROS, and in 1960, the American Meteorological Society presented Kellogg and Greenfield with a special award.

The RAND Corporation was an ideal place for Kellogg’s and Greenfield’s work. The organization was formed in 1948 when the U. S. Air Force, previously the Army Air Forces, became a separate branch of the armed services. It grew from Project RAND, which the Douglas Cor­poration set up immediately after the Second World War to evaluate advanced technology for the Army Air Forces’ missions. RAND made its
first analysis of the technical feasibility of satellites for the Army Air Forces in 1946.

In 1948, the government assigned responsibility for satellites to the Air Force, which appointed RAND to manage the work. RAND subcon­tracted studies on guidance, stabilization, electronic reliability, communica­tions, space reconnaissance systems, and space power systems to companies such as Westinghouse, Bendix, and Allis Chalmers. In 1951, RAND pub­lished classified studies incorporating industry’s and its own work. These studies analyzed potential missions as well as engineering design, the polit­ical implications of the technology, and the potential of satellites for sci­ence. Later, when the U. S. space program got seriously underway follow­ing the launch of the first two Sputniks, the content of these reports would be incorporated into the early classified satellite programs, such as the Dis­coverer series (on which APL had a transmitter).

RAND’s main conclusion in 1951, however, was that satellites had potential as observation platforms for reconnaissance. In the same year, Kellogg and Greenfield completed a study on the feasibility of satellites— referred to as satellite missiles or satellite vehicles—for “weather reconnais­sance.”

Weather and reconnaissance satellites proved to be a hard sell, but of the two it was reconnaissance satellites that first won the administration’s backing. In winning that backing, reconnaissance satellites precipitated, by the tortuous paths that characterized the Cold War, the U. S. space pro­gram, including—eventually—meteorology satellites.

Initial opposition to reconnaissance satellites focused on their techni­cal limitations. In those days photography from high altitudes offered spa­tial resolutions of the order of hundreds of feet, knowledge obtained from high-altitude rocket photographs (thirty, forty-five, and sixty-five miles), which the Navy had been the first to take. This resolution was much poorer than what could be obtained from cameras on aircraft, and those who had developed expertise in photo interpretation during World War II were scornful of the technology’s capabilities.

But prompted by the need for better intelligence, both to prevent surprise attacks and to monitor arms-control agreements, the Eisenhower administration proposed low-level funding for a reconnaissance satellite program in fiscal year 1956 (that is, for funds available from October 1955). The money authorized was as follows: fiscal year 1956—$4.7 mil­lion; 1957—$13.9 million; 1958—$65.8 million.

Air Force historian R. Cargill Hall says that the critical factors in winning backing for the development of a reconnaissance satellite were that satellites offered a photographic platform that could not be seen by the naked eye or detected by radar sensors, and if satellites were detected, they would be too far away to be shot down.

These were not advantages that the Soviets were likely to appreciate. When President Eisenhower suggested in 1955 that there be an “open skies” policy, allowing overflight of Soviet and U. S. territory in order to ver­ify disarmament agreements, Khrushchev had dismissed the idea, calling it “licensed espionage.” Khrushchev viewed the policy as an attempt to gather information on potential military targets, and that, indeed, was one of its purposes. So the prospect of American spy satellites could be guaranteed to provoke Soviet animosity. For these and other reasons, argued the historian Walter McDougall, the Eisenhower administration wanted a civilian satel­lite launched first to establish the precedent of the freedom of space. In an article in Prologue, Quarterly of the National Archives (spring 1996), Cargill Hall confirms this view, arguing that the IGY enterprise effectively was made into a stalking horse for military reconnaissance satellites.

It was not a cheap stalking horse—nearly $20 million initially for the satellites alone—but it was effective. The program played host to enough political tensions, technical difficulties, protests from the Army, and criti­cisms of the Eisenhower administration’s Vanguard decision to keep the eyes of the world focussed on the IGY. Thus it was that IGY became the forum in which the first satellites were launched.

To the disappointment of Kellogg and Greenfield, weather satellites languished during the discussions about reconnaissance. Even though the resolution was poor, satellite images, they believed, offered something of potentially great importance to meteorologists—a “synoptic” picture, assembled from several individual photographs, showing cloud patterns over a large expanse of the earth at one time.

The proposal Kellogg and Greenfield made in 1951 had to wait seven years until the Advanced Research Projects Agency was prompted by Sputnik into finding applications for satellites. William Kellogg was then appointed to head an ARPA panel drawing up specifications for a meteorology satellite.

Kellogg turned to his and Greenfield’s early ideas, which, as in the case of reconnaissance satellites, owed much to rockets and high-altitude balloons. Images from rockets enabled meteorologists to begin developing analytical techniques that made sense of the bird’s-eye view of the Earth, while balloons served as test beds for meteorological instruments and cameras. The balloons could carry more weight aloft than could early satellites.

In the late 1940s, high-altitude balloons were already a versatile technology for a variety of applications. William Kellogg had worked on a project for the U. S. Atomic Energy Commission investigating their potential for monitoring the dispersion of radioactive particles from atomic tests; the Japanese, as we saw, had used them to carry incendiary bombs; and the Photo Reconnaissance Laboratory at the Wright Field in Ohio established that balloons made a stable platform for aerial photogra­phy.

At about the same time, in January 1949, the Bulletin of the American Meteorological Society published an article by Major D. L. Crowson, “Cloud Observations from Rockets.” Crowson suggested that even low-resolution imagery from high altitudes would improve weather forecasting.

Given this background, it is not surprising that Kellogg and Green­field decided to pursue the idea of using high-altitude balloons first for photo reconnaissance and then for meteorological research. This work gave them the information they needed about optical systems for their 1951 report on weather reconnaissance from a satellite. What they wrote was by no means a detailed engineering proposal, but it tackled concep­tually for the first time the elements of a meteorological satellite carrying a camera operating in the visible portion of the electromagnetic spec­trum.

They asked, Can enough be seen from an altitude of 350 miles to enable intelligent, usable weather observations to be made, and what can be determined from these observations?

From analyses of photographs taken by rockets, they decided that a resolution of five hundred feet was necessary if all the useful cloud struc­tures were to be identified.

They assumed that the camera would mechanically scan to build a photograph of a wide enough area—a 350-mile swath around the sub­satellite point—to be of use. Once they had specified the minimum reso­lution, they discussed the aperture, illumination, exposure time, and focal- length-to-aperture ratio needed to achieve a given contrast. The satellite, of course, would not be taking carefully posed and cunningly lit pho­tographs but would have to operate in whatever conditions nature de­creed. So the camera and optical system had to be chosen to provide usable photographs in a variety of conditions.

Photographs, they recognized, would not yield the quantitative information that meteorologists needed. It would be impossible to do more than make intelligent guesses at temperatures and pressures. But in those days, before numerical weather prediction, these limitations did not seem as great an obstacle as they would shortly become. Thus, unknow­ingly, Kellogg and Greenfield put their finger on what was to be the mam problem in winning acceptance for satellite meteorology. Analysts, they wrote, would have to make the most of the visible aspects of the weather in building their weather charts; clouds, being the most visible aspect of the weather, would be important.

Rocket photographs had already given an inkling of the inherent problems. In pictures taken by cameras on Thor and Atlas, it was difficult to tell whether areas of uniform greyness were cirrus clouds (wispy, high- altitude clouds formed of ice particles), tropospheric haze, or an artifact of the optical system resulting from the wavelengths accepted. All was not bad news, however, because more cloud patterns had been apparent in rocket photographs than had been expected.

In an attempt to get a feel for how accurately one might forecast weather from satellite photographs, Kellogg and Greenfield had estimated the synoptic situation from photographs taken during three rocket flights and had then made a forecast and compared it with records of actual weather on the day in question. Encouraged by the results, the two con­cluded that “combined with both theoretical knowledge and that gained through experience, an accurate cloud analysis can produce surprisingly good results.”

By the end of 1959, anticipating the launch of TIROS in 1960, lead­ing researchers met in Washington to discuss cloud research, a field known as nephology. Harry Wexler, then chief scientist of the Weather Bureau, pointed out that until the late nineteenth century, clouds had been almost the only source of information about conditions in the upper atmosphere, but that with the advent of balloon soundings, interest in nephology had declined. (Now clouds are recognized to be of crucial importance in meteorology, particularly in climate studies, and meteorologists are asking such basic questions as “What is a cloud?”)

Sig Fritz, who worked for Wexler, spoke of the strong sense researchers had that they would not know how to interpret cloud pho­tographs. This was a problem that Kellogg and Greenfield had foreseen in their 1951 report, and they had recommended that in preparation for satellite images, meteorologists build a comprehensive atlas of clouds as seen from above.

TIROS would soon begin that process.

The conception and birth of the TIROS satellites were difficult. First, in late 1957, the secretary of defense, Neil McElroy, agreed that a new agency—the Advanced Research Projects Agency (ARPA)—would have responsibility for key defense research and development projects. ARPA officially opened its doors on February 7, 1958, and weather satellites became one of its projects.

Immediately, Kellogg was asked to define the specifications for a weather satellite, which he did with the help of people like Dave Johnson. In the meantime, RCA had submitted a proposal to the Army Ballistic Missile Agency for a reconnaissance satellite as part of the Redstone mis­sile program.[11] The Department of Defense decided that the images from this satellite would not be good enough for intelligence gathering, and it therefore became TIROS, modified to incorporate the conclusions from Kellogg’s group at ARPA. ARPA managed the TIROS project until the National Aeronautics and Space Administration took it over in October 1958.

The newly formed NASA was a powerful organization, and it could expropriate groups and organizations. One of the groups that the agency wanted was that headed by Bill Stroud, of the Army’s Signal Corps of Engineers, which was working on a camera for the IGY’s cloud-cover experiment. After some altercations with military bureaucrats, Stroud was able to transfer to NASA in 1959, and he headed the agency’s meteorology branch at the Goddard Space Flight Center.

So TIROS had its roots in a spy satellite proposed to the Army by RCA but became a weather satellite managed first by ARPA and then by Stroud’s group at NASA.

The satellite’s optics had a field of view four hundred miles on a side. It carried a small Vidicon TV tube, selected because of its light weight.

The images recorded were poor because the camera’s electron-beam scan was not well controlled. Also, recalled Verner Suomi, who soon became involved with the TIROS program, “We didn’t know what the devil the damned thing was looking at. There were some problems as to what time the pictures were taken, and the spacecraft was spinning like a top. Where the devil was north? That caused major problems.”

Sean Twomi, of the Weather Bureau, soon identified the cause of one of the problems. The spacecraft needed to spin to remain stable in orbit, but the spin axis was tilting because of interactions between the spacecraft’s electrical systems and the earth’s magnetic field. Thus it was difficult to know where the cameras were pointing. Once Twomi had identified the problem, the Air Force developed magnetic stabilization to control the orientation of the spin axis, though it was some time before the engineer­ing problems of orientation and stabilization of weather satellites were fully solved.

Despite such difficulties, those involved, like the engineers and scien­tists at the Applied Physics Laboratory, had an overwhelming sense of being pioneers. Bob Ohckers, an engineer who worked on the TIROS program at RCA and later moved to Suomi’s lab at the University ofWis – consin, recalled, “In those days, there were no cutesy requirements, no quality control or oversight. Everything was experimental. If we had a fail­ure, we would try to keep it from the contractor, particularly from Thomas Haig, who headed the Air Force’s weather satellite program (Haig also joined the University ofWisconsin, where he and Suomi spent some time feuding). Haig would try to ferret out what was going on. We were told to tell him nothing. The whole group was working with rolled up sleeves and screwdrivers.”

Despite the limitations of the TIROS satellites, both in terms of the data they collected and of the analytical techniques available for data processing and interpretation, the first images returned to Earth were tantalizing.

In 1964 Suomi gave a lecture to children at a local school. A copy of the speech, recorded by some attentive listener, was in an old filing cabinet in the basement of the building where Suomi worked at the University ol Wisconsin. He told them how unsophisticated and crude the satellites were. But he also told them that during one orbit of TIROS they had identified two hurricanes in the southern hemisphere before they had been spotted by ships or weather stations. He showed them bright areas of cloud, telling them that this meant that the clouds were thick and high and represented an enormous thunderstorm, but that they only learned these things after the event. “We have much to learn about how to apply these pictures. The future depends not on the hardware, not on the gadgets, not on the software, but on individuals applying their knowledge to this very challenging problem [of interpretation].”

With the formation of NASA, the defense and civilian meteorology satellite programs went their separate ways. Until 1965, the Defense Mete­orology Satellite Program (DMSP) was one of a suite of satellites con­trolled ultimately by the CIA—an indication of how intertwined the mis­sions of reconnaissance and meteorology satellites were. Then, in 1965, control of funding for DMSP was transferred to the Air Force. DMSP remained secret until John McLucas, the under secretary for the Air Force, made the program public in 1973.

Since 1958, when the two programs diverged, the civilian and defense meteorological worlds have resisted all political efforts to reunite

them. Meetings about a merger between the two were, says one partici­pant, like arms control negotiations of old, where people developed their arguments as to why arms control was not possible.

Suomi, who like many of his colleagues worked on both sides of the fence, recalled that he would see the same work being done twice. “Part of my activities here [at the University of Wisconsin], which were classified

then, was to put the heat budget experiment on a military satellite. The thing that was interesting was that many of the things that the civilian pro­gram utilized were actually developed for the military. What interested me as a party to both was that I saw one part of the RCA building which was classified and I saw another part of the building not classified. Both parts were classed as development, but really they were only one development. Someone made a lot of money on that.”

Johnson, too, saw duplication, but said that interactions between the two worlds could work well when individuals in the military program were cooperative. On one occasion Johnson wanted to fly a new tube, and the DMSP allowed one of its own tubes to be replaced by Johnson’s. Many of the DMSP people would also, says Johnson, do what they could to move expertise from the “black” world. But technology was not always transferred. Suomi recalled that DMSP effectively equipped spacecraft

with “exposure meters” so that they could photograph clouds by moon­light in the visible part of the spectrum. Asked if such technology would have helped civilian weather satellites and why it was not transferred, he answered, “Yes” and “I don’t know.”

Though declassification may clarify some of these questions, it is unlikely to revise substantially the pioneering role that Verner Suomi played.