Overview
Project Cirrus was a joint weather-modification programme run by the General Electric Research Laboratory with the US Army Signal Corps, the Office of Naval Research and the US Air Force, and it ran from 1947 to 1952. It is not a disputed programme. It was publicly announced, its results were published in the scientific literature, and the men who ran it are named.
What keeps it in circulation is a single afternoon. On 13 October 1947, aircraft working for the project flew out over the Atlantic and dropped crushed dry ice into a hurricane. Two days later that hurricane, having reversed its course, came ashore at Savannah, Georgia. The director of the laboratory then told the public he was almost certain the experiment had caused the change.
Where the idea came from
The programme grew out of a laboratory accident, and the accident is well recorded.
In 1946, Vincent Schaefer, working at GE's laboratory in Schenectady, New York, was trying to produce a supercooled cloud in a chest freezer — water held below freezing that has not yet turned to ice. To cool the box faster he dropped in a piece of dry ice, and the cloud inside instantly filled with ice crystals. He had found that solid carbon dioxide would nucleate ice in a supercooled cloud, and that the effect needed almost nothing to trigger it.
The first field test followed on 13 November 1946. A GE pilot, Curtis Talbot, took Schaefer to fourteen thousand feet about thirty miles east of Schenectady and released three pounds of dry ice into a cloud. Schaefer recorded seeing "long streamers of snow falling from the base of the cloud." A colleague, Bernard Vonnegut, shortly afterwards found that silver iodide would do the same thing, and because it could be released from the ground rather than from an aircraft it became the standard material for the decades of weather modification that followed.
Project Cirrus was the government-funded programme those two discoveries produced. Its purpose was to find out what else could be done to a cloud.
The hurricane
By the autumn of 1947 the project wanted to try a hurricane, and the storm that presented itself was the one the Air Force had designated King.
King crossed southern Florida on 12 October, dropping enough rain to cause damaging floods that lasted for days, and then moved back out over the Atlantic and headed away from the coast. That was precisely what made it attractive: the storm was no longer over land, it was travelling out to sea, and whatever happened to it afterwards should not have reached anybody.
On the afternoon of the 13th, three aircraft left Mobile, Alabama. The lead B-17 flew at cloud-top height and did the seeding, with large blocks of dry ice fed through a crusher in the aircraft and dropped from the belly into the cloud below. A second B-17 followed half a mile back and three thousand feet higher to photograph what happened. A third carried Schaefer, directing the other two and watching for changes.
They did not go into the eye or the heavy rainbands — the aircraft had no equipment good enough to find their way back out — so they worked the outer clouds, about four hundred and fifteen miles east of Jacksonville. They made a single run of roughly a hundred miles lasting half an hour, releasing eighty pounds of dry ice, then turned back to look at what it had done. They then dropped fifty pounds into one large cumulus tower, and fifty pounds again, and circled to watch. The cloud deck beneath the first run began to break apart. The seeded tower kept growing. The crew logged a "pronounced modification of the cloud deck seeded," and the aircraft went home.
The turn
The scientists flew out the next day to look at the storm and could not find it where it was supposed to be.
When they located the centre it was nearly a hundred miles west of the forecast position. The hurricane had made a turn of about 135 degrees to the left and was now travelling due west, towards the coast, and it was getting stronger. On the afternoon of 15 October it came ashore at Savannah. One person was killed in the storm surge, and the damage across Georgia and South Carolina came to around two million dollars.
The reaction was immediate and it was angry. People in the affected states believed the scientists had turned the storm onto them, and lawsuits were threatened against General Electric and against the government.
The statement that made it worse
What raised the episode from an accusation to a documented controversy is that the man in charge agreed with the accusation.
Irving Langmuir, the Nobel laureate who headed GE's laboratory, said publicly that he was "99% sure" the hurricane had changed course because of the seeding. He was the senior scientist on the programme, he was speaking about his own experiment, and he was saying the thing the public already feared.
The counter-argument came from the Weather Bureau. Its chief, Francis Reichelderfer, took the position that hurricanes do this on their own, and set three of his forecasters to find a historical storm that had made a comparable turn without anyone seeding it. They found one, and the case was published. The threatened lawsuits faded, and the question of what had actually moved King was left where it has stayed: a storm that turned after an intervention, with no way to run the experiment again without it.
What happened to the programme
Project Cirrus continued until 1952, but the hurricane work stopped.
The immediate effect was institutional caution. For years afterwards, hurricanes and weather modification were not discussed together in public by anyone who wanted to keep working on either. It took eleven years before the National Hurricane Research Project tried modest seeding equipment tests in a storm, and they were kept deliberately quiet until it was clear the hurricane was not going to do anything dramatic.
The formal successor came in 1962, when the Weather Bureau and the Department of Defense agreed on Project Stormfury, which ran until 1983 and tried to weaken hurricanes by seeding convection outside the eyewall with silver iodide. Four hurricanes were modified on eight days; on four of those days the winds dropped by between ten and thirty per cent, which looked promising at the time.
Those results did not survive the 1980s. Closer observation of unmodified hurricanes showed that the double eyewalls Stormfury had read as a success happen routinely without anyone seeding anything, and that hurricanes contain little of the supercooled water that cloud seeding acts on at all. NOAA's Hurricane Research Division has since put it flatly: there is no sound physical hypothesis for the modification of hurricanes. The mechanism the whole effort depended on was not present in the clouds it was aimed at.
Why it became a reference point
Project Cirrus is cited constantly in later weather-control claims, and the reason is that it does not require any of the usual steps. There is no need to argue that a secret programme existed, that it was concealed, or that officials lied.
The programme was real and was funded by three branches of the US military. The seeding happened on a known date. The hurricane did turn, and it did reach a city, and a person did die. The scientist responsible did say he thought the experiment caused it. The agency that disputed him did so by commissioning a study with a specific purpose, after liability had been raised.
Every one of those is on the record, and together they make a strong starting position for an argument the record does not actually settle. What the documents establish is that an experiment took place and that a storm changed course. Whether the first produced the second is the one part nobody has been able to demonstrate, in either direction, and eighty years of subsequent work suggests the quantity of dry ice involved was far too small to move a system of that size.
The claim that the state can steer a storm at a city is older than the satellite era, and this is where it acquired its best piece of evidence.