Fixing the Sky

by James Rodger Fleming

  Project Cirrus

  In February 1947, General Electric research director Suits hurriedly called a halt to outdoor experimentation on cloud seeding and instructed Langmuir’s team to serve only as advisers on Project Cirrus, a new classified cloud-seeding effort to be conducted by the U.S. Army Signal Corps, the Office of Naval Research, and the U.S. Air Force. As stated in the GE contract, the general purposes of the project were “research study of cloud particles and cloud modifications” by seeding, including investigations of liquid water content, particle sizes and distribution, and vertical cloud development.38 They were searching for fundamental knowledge of cloud physics and chemistry to improve operational forecasting as well as practical techniques of cloud modification for military purposes or possible economic development.39 An important clause in the contract further stipulated that “the entire flight program shall be conducted by the government, using exclusively government personnel and equipment, and shall be under the exclusive control of such government personnel.” Suits notified his staff that “it is essential that all of the GE employees who are working on the project refrain from asserting any control or direction over the flight program. The GE research laboratory responsibility is confined strictly to laboratory work and reports.”40

  GE argued that the whole matter properly belonged to the government, and that the government, by suitable legislation, should both regulate the inducing of rainfall and indemnify for loss any contractor acting on the government’s behalf—especially themselves. Secretary of Defense James Forrestal asked Congress for a law “to protect contractors engaged in cloud modification experiments against claims for damages by third parties,”41 but no such legislation was forthcoming. The Harvard Law School Record reported:

  Today “Project Cirrus” has an annual budget of $750,000 from military and naval funds because of its war implications—bogging down enemy troops in snow and rain, clearing airfields of fog at lowest cost, and infecting induced storms with bacteriological and radiological materials. The Battle of the Bulge, in which the Nazis mobilized and attacked under supercooled fog, could have been much altered by a few pounds of dry ice.42

  Between 1947 and 1952, Project Cirrus conducted about 250 experiments involving modification of cold cirrus and stratus clouds, warm and cold cumulus clouds, periodic seeding, forest fire suppression, and a notable attempt to modify a hurricane. Researchers in the project developed a suite of modern techniques applicable to cloud physics, including instruments for measuring temperatures and cloud properties in flight, collecting cloud droplets and ice crystals, and generating artificial nuclei.43 Military aircraft (a B-17, later a B-29, and eventually as many as six planes) equipped with seeding devices, new instrumentation, and camera equipment operated over a 1,000-square-mile restricted flight area just north of the Schenectady airport, where the team was based. Under the auspices of Project Cirrus, Langmuir consulted with cloud seeders in Central America and corresponded with cloud seeders in Hawaii who were seeking to generate rainfall from warm convective clouds. This stimulated Langmuir’s thinking about possible chain reactions in cumulus clouds seeded by as little as a single drop of water. Although the Project Cirrus staff collected and analyzed mountains of photographic and other data, the response of the atmosphere to seeding was erratic and the researchers could not obtain any definitive measures of the efficacy of artificial nucleating agents. The results from several experimental runs were spectacular, however, and the Department of Defense decided to expand the work of Project Cirrus to include rain enhancement experiments in New Mexico, forest fire suppression trials in New England, liquid water seeding of warm clouds in Puerto Rico, and hurricane modification in the Atlantic Ocean.44

  One parallel study, the joint Air Force–Weather Bureau Cloud Physics Research Project, found that seeding did indeed produce striking visual changes in clouds, including dissipation of cold stratus decks. However, experiments with clouds over Ohio in 1948 and over California and the Gulf states in 1949 led the researchers to conclude that cloud seeding could not initiate self-propagating storms or relieve drought. The weather bureau spent $85,000 on the project in 1948 and $100,000 in 1949, with the air force supplying aircraft, personnel, and ground radar facilities.45

  Hurricane King

  In October 1947, GE announced that Project Cirrus would be intercepting a hurricane, not to “bust” it but to experiment on the effects of seeding with dry ice on a portion of a storm. Atlantic tropical storm number eight, unofficially dubbed Hurricane King, had just made a devastating pass over southern Florida and was churning in the Atlantic Ocean about 400 miles northeast of Orlando. It was expected to head farther out to sea. On October 13, the Project Cirrus team, led by navy lieutenant commander Daniel Rex and accompanied by Schaefer, bombed the heart of the storm with 80 pounds of dry ice and dropped 100 pounds more into two embedded convective towers.

  The newspapers initially reported that the task force had “attacked” the storm in a “hurricane-busting” effort to reduce its winds or redirect it. It was reported in the press as “history’s first assault by man on a tropical storm,” an experiment with energies of nature far greater than those unleashed by the atomic bomb.46 The official results were classified as military secrets, and Schaefer told the press that he was “not allowed to say” whether the seeding had had any visual effects.47 Commander Rex’s official report, not yet released, claimed a pronounced modification of the cloud deck that had been seeded. What happened after that, according to Langmuir, “nobody knows,” since Hurricane King made a “hairpin” turn and headed west, smashing into the coast along the Georgia–South Carolina border near Savannah (figure 5.3). In Charleston, a tree fell, killing one person, and the storm caused more than $23 million in damage during its second landfall. 48 A letter in a St. Petersburg newspaper from J. M. Enders and addressed to GE research director Suits placed the blame for the devastation on “the weather tinkers of your lab” and pointed out that the people of Savannah were not so sure it was a coincidence. In fact, they were “pretty sore at the army and navy for fooling around with the hurricane.”49

  5.3 Project Cirrus hurricane-seeding experiment and the subsequent path of Hurricane King in 1947 (solid line), compared with the path of a 1906 hurricane (dashed line) that also had turned suddenly. A retrospective study by the weather bureau showed that upper-level steering currents, not seeding with dry ice, had likely caused the storm to veer suddenly. (SCHAEFER PAPERS)

  No one held the “hurricane busters” officially liable, but that would certainly not be the case today. The storm’s unexpected turnaround following—if not necessarily because of—seeding dampened GE’s hopes of making grandiose claims about storm control. Schaefer participated in a press conference at which evading questions was the order of the day, and he wrote in his official report: “Change in plans of the publicity angles to the project caused considerable delay and should be completely eliminated. This should be done by the assignment of a [public relations officer] to the project if it’s again tried.”50 An unrepentant Langmuir admitted, “The main thing we learned from this flight is that we need to know enormously more than we do at present about hurricanes.”51 Langmuir was already looking ahead to future hurricane seasons—he hoped that the Project Cirrus team could intercept hurricanes far out at sea, fly multiple tracks through them, “and see if we cannot, by seeding them, in some way modify or shift their positions.... The stakes are large and, with increased knowledge, I think we should be able to abolish the evil effects of these hurricanes” (185).

  Six decades later, the case of Hurricane King might serve as a warning to the Department of Homeland Security, which, as of 2008, wants to fund a new wave of research aimed at weakening the strength of tropical storms and steering them “off course.” But, of course, hurricanes do not run on tracks or on a schedule, so everyone damaged by a modified hurricane could sue for damages—unless the government tried to place an embargo on such lawsuits.52

  Silver Iodi
de

  The exciting news from GE about weather control took another step in January 1947 when physical chemist Bernard Vonnegut discovered that molecules of silver iodide act as artificial nuclei and can “fool” cloud water droplets into crystallizing.53 During World War II, Vonnegut worked in the Department of Chemical Engineering at MIT on projects related to gas warfare and with the Department of Meteorology on problems of aircraft icing. He moved to GE in 1945 and worked closely with Langmuir and Schaefer. His brother, the famous writer Kurt Vonnegut, also worked at GE as a publicist.

  Following Schaefer’s cold box discovery, Langmuir asked Vonnegut to do quantitative work “on the number of ice crystals produced by dry ice.” This led Vonnegut to search for other agents that might initiate ice phase processes in a cloud. As he told the story five years later, “It occurred to me that if I could get something that was awfully close to ice in its crystal structure that might do the job, and I looked up in the handbook to find out what substances were close. I came across ... lead iodide, antimony, and silver iodide.”54 Powdered lead iodide produced “a reasonable number of crystals” in Schaefer’s cold box, a phenomenon that Schaefer attributed to the hexagonal shape of the molecules, but still they could not get a good result: “Well, I couldn’t figure this out; so I was just puttering around and I decided to see just what happened when I put metal smoke in there [from silver]. I was amazed—the ice box was just swimming with ice crystals—colossal numbers.... Then I remembered silver iodide and made smoke.... Vaporizing silver iodide worked like a charm.” Vonnegut’s laboratory notebook indicates that he had identified lead iodide as an artificial nucleating agent by November 6, 1946. After numerous trials, he finally got iodine vapor and metallic silver to work on November 14: “Hallelujah! the nucleation was even more wonderful.” By November 18, just four days after Schaefer’s outdoor experiment, Vonnegut had found out that it was the silver iodide that did the job. Because of its hexagonal structure, silver iodide imitates ice condensation nuclei, causing “explosive ice growth” in supercooled clouds.55

  Vonnegut soon started seeding experiments with ground-based silver iodide generators, but found it hard to tell where the smoke was going and what effect it was having. He did some inconclusive aircraft tests in December 1947, but “felt” that the experiments he conducted with GE meteorologist Ray Falconer with a silver iodide generator on the summit of Mount Washington were the most satisfying : “I feel darn sure we got some nice results ... we caused quite a nice snow squall downwind from the generator on air as it goes up over the mountain.”56 He had yet to develop reliable techniques for following the particles and measuring their concentrations. The General Electric News Bureau, however, was quick to claim credit:

  If generators can be used on the ground to introduce silver iodide or other foreign particle nuclei into huge masses of air, it might be possible to alter the nature of the general cloud formation over the northern part of the United States during winter.... It would prevent all ice storms, all storms of freezing rain, and icing conditions in clouds. The amount of heat absorbed by sunlight would be changed. It should be possible to change the average temperature of some regions during winter months.57

  Vonnegut, who was less sanguine, pointed out that silver iodide has its own problems. It is persistent in the environment and can activate long after its release in clouds of proper temperature; dry ice, though, works immediately and then sublimates. He recalled in an interview, “This is bad, I think, for commercial cloud-seeding operations, because I think they’re playing with fire releasing this stuff all over the place and I think it’s a shame they haven’t shown any sense of public responsibility particularly when they deny it has any large scale effect ... to stink up the atmosphere for hundreds or thousands of miles down wind producing God knows what effect is a dangerous thing.”58

  Nevertheless, Langmuir touted chemical seeding agents as being superior to natural ice nuclei because they act at higher temperatures, they do not melt or evaporate, and they can be spewed into the atmosphere over widespread areas to remain active until it snows. He echoed the GE News Bureau in making unsubstantiated claims that the chemical might eliminate severe aircraft icing, suppress hailstorms, and perhaps, since by his estimates only 200 pounds of silver iodide would be needed to seed the entire atmosphere of the United States, could result in large-scale weather or even climatic changes.59

  The New Mexico Seedings

  Project Cirrus was operating in Socorro, New Mexico, in July 1949, with Vonnegut running a test burner on the ground while the military air crews, with Langmuir and Schaefer as observers, seeded clouds aloft with dry ice. One day, one of Vonnegut’s cumulus clouds “really whooped it up, and the first thing I knew there was lightning, and it was all very exciting, and I thought, ‘Gee, I wonder if I’m responsible for it.’”60 When his colleagues returned that evening with no results to report, Vonnegut ventured the suggestion that perhaps his ground generators had glaciated all the clouds, so the dry ice seeding planes could not find any for their experiments.

  Noting that widespread rains were reported downwind on this day, Langmuir ordered seeding to be done periodically, once a week for eighty-two weeks, from December 1949 to July 1951. Then, even before the data were collected and other possibilities explored, he proceeded to make the outrageous claim that large-scale seven-day periodicities in the nation’s weather were being caused by Vonnegut’s single ground-based silver iodide generator located in New Mexico. Langmuir supported his claim by noting that the Midwest and East were moister than normal during this period, while the Southwest was drier than normal—conditions that occur naturally on a regular basis, although he did have to admit, sheepishly, that in the weeks when the generators were not operating, “rains were about forty percent greater than the other weeks.”61 Nevertheless, Langmuir went so far as to claim that severe flooding in the Midwest and the Ohio Valley, accompanied by widespread property damage and loss of life, was the result of these experiments. He apparently “proved” his result using unconventional statistical methods of his own devising.

  In his insightful autobiographical memoir, Sverre Petterssen, a leader in the field of weather analysis and forecasting, reviewed his involvement with Langmuir in the 1950s. Petterssen was trained in the Bergen School of meteorology, chaired the Department of Meteorology at MIT, and served in Norwegian uniform in World War II, preparing forecasts for bombing raids by the British Royal Air Force, the Anzio landing, and, notably, the D-day invasion of Normandy. After World War II, Petterssen served as head of the Norwegian Forecasting Service, scientific director for the U.S. Air Force Weather Service, and professor and chair of the Department of Meteorology (later Geophysics) at the University of Chicago. He explained the situation:

  About 1947 Irving Langmuir, a Nobel Prize laureate in chemistry, and his group working at the General Electric Laboratories had discovered that silver iodide had a structure similar to small ice crystals. Since natural clouds, even at very low temperatures, are generally deficient in ice crystals while silver iodide can readily be produced, it seemed possible to supply silver iodide dust to cold clouds, hoping that the clouds might be “fooled into believing” that natural ice crystals were present. Thus, if the clouds could be so misled (and few doubted it) weather modification (or control) would not be much of a problem.

  Langmuir was unlucky and became a victim of one of the many pitfalls that nature so generously provides for scientists who venture too far outside their own field of specialization. Though a leading authority on the chemistry of crystal points and surfaces, a philosopher, and a polyhistor in general science, Langmuir did not appreciate the complexity of meteorology as a science. In the atmosphere, processes of vastly different spatial scales and life spans exist together and interact; impulses and energy are shuttled through the whole spectrum of phenomena—all the way from molecular processes to global circulations and the changes in the atmosphere as a whole. No chemist, physicist, or mathematician who has no
t lived with and learned to understand this peculiar nature of meteorology can pass valid judgment on how the atmosphere will react if one interferes with the details of the natural processes. Moreover, to determine whether or not the atmosphere has responded to outside interference, it is necessary to predict what would have happened had it been left alone.

  As I have just said, Langmuir was unlucky. For no profound reason he had left a silver-iodide generator somewhere in New Mexico and made arrangements with a local person to “burn” the generator on a weekly schedule. Using a set of readily available weather reports, Langmuir found that the rainfall had begun to vary in a weekly rhythm. The amazing thing was that the response was not just local; it was nationwide and might well be of hemispheric proportions. Langmuir, and many with him, concluded that the weekly injection of silver iodide from a single generator in New Mexico had excited a hitherto undiscovered natural rhythm of the atmosphere, with the result that the rainfall had yielded to the will of man. ... In his mind, and in the minds of many others, there was but little doubt that the weather processes could be intensified or repressed to suit human needs.62

  Langmuir was unable to accept the criticism of Petterssen or the analysis of weather bureau statistician Glenn Brier that the atmosphere frequently exhibits a natural seven-day periodicity.63

  After the New Mexico incident, Suits again warned Langmuir that his field experiments and unsupported claims might put Project Cirrus at risk and expose the lab to litigation. He pointed out that Schaefer and Vonnegut were “a great deal less certain” about the interpretation of the New Mexico results than he was, and that ground-based seeding would again raise legal questions for General Electric: “If the [cloud seeding] program develops in such a direction as to subject the Company to serious hazards from a liability standpoint, it may very well become impossible for us to continue with this work.” In a long letter to Langmuir, with carbon copies to Schaefer and Vonnegut, Suits reminded the team that “there has been no recent change in the law which makes it less necessary at present for General Electric personnel to be cognizant of the hazards from the standpoint of legal liability than when the agreement referred to above was reached.” Suits again reminded the team that GE employees were to serve only as advisers to the government: “GE personnel must not engage directly or indirectly in seeding experiments which might lead to harmful weather phenomena. They may engage in laboratory experiments which they consider advisable and in very small scale weather experiments for confirming laboratory tests with actual meteorological conditions.” Suits could not approve their publications that reported the results of large-scale modification. He issued a similar embargo on technical talks, claiming that GE was doing the experiments for the sake of humankind and was earning no profits from the activity: “I do not believe that our obligation extends to the taking of exceptional risks of damage suits as a result of any work which we may do in this field.”64