Can cloud seeding save California?

Will Bates
11 min readOct 22, 2020


With the state looking at spending billions on fire, it’s worth a shot.

We’ve seen the photos of the apocalyptic skies — and the Twitter jokes:

“Everybody complains about the weather,” Charles Dudley Warner wrote in 1900. Californians have had a lot to complain about this year. Warner finished his quip with, “But nobody does anything about it.”

Actually, California has a long history of both research into and the actual practice of weather modification going back some 70 years.

California’s current cloud seeding projects increase its water supply by 4 percent a year.

According to 2014 figures from state’s Department of Water Resources, cloud seeding increases the state’s water supply by at least 4 percent a year. Cloud seeding turns out to be an extraordinarily cheap source of extra water. It’s environmentally preferable to pumping the aquifers; orders of magnitude less expensive than desalinization; and more cost-effective than conservation measures.

Yet what California does now, called orographic cloud seeding, is done in winter over mountain watersheds with the goal of increasing runoff for hydropower, municipal supplies, and irrigation, and is financed by local water districts, utilities like PG&E, and even some ski resorts.

The obvious question is: could California do more cloud seeding to get rain where it needs it — on the dry forest fuel load left by a century of misguided fire-suppression policies? Could it get that rain when it needs it, in late spring, before the fire season begins?

Record heat…

The exceptional thing about California’s 2020 fire season may be how routine “worst-ever” is starting to sound. Pulled by the unstoppable locomotive of global warming, each new year becomes a contender for a new worst year. July 2020 was the hottest month in California’s recorded history. The state’s five warmest years occurred in between 2014 and 2018. Worldwide, 2019 was the second warmest year since record-keeping began in 1880. Thus far, 2020 has an excellent shot at the hottest-ever title.

While short-term variations in can always provide a break in the weather, the underlying trend makes it clear things will not improve anytime soon. Even if we could wave a magic wand and end CO2 emissions today, the greenhouse gas in the atmosphere right now will take — estimates vary — on the order of 50 to 200 years to dissipate, if then. Nor is a long view of much comfort. The Paleocene–Eocene Thermal Maximum (PETM), in which volcanoes put a comparable amount of carbon dioxide into the atmosphere, lasted 200,000 years. Absent a technological miracle, global warming is not going away.

…and drought.

Of the many consequences of warming, the deadliest of these is drought. Or so concludes historian Brian Fagan in his 2008 book The Great Warming, a study of a climate anomaly between 950 and 1250 AD.

Unlike hurricanes and stranded polar bears, drought is not telegenic — until it produces fire.

Unlike hurricanes and stranded polar bears, drought is not telegenic — until it produces fire. While California has always suffered from periodic droughts — it may be falsely reassuring to speak of swings around the current trend line as “periods” at all — the last decade in California has been something else.

February 2020, normally a rainy month, was the driest in recorded history. Near the start of the fire season in June, 60% percent of California was in a “dry” condition, as was 96% of Oregon. Then came July with its record-break heat.

The prospects for 2021 are not better. Sea-surface temperatures in the central and eastern tropical Pacific Ocean, early indicators for winter El Niño and La Niña conditions, suggest a 75 percent likelihood 2021 will be a La Niña year — and a dry winter for the Pacific Coast and American Southwest.

The Forgotten History

California’s first experiments with weather modification date back to the early 1950s, not long after the discovery of the technique for cloud seeding at General Electric Research Laboratory in upstate New York in 1946.

The lab had studied the problem of airplane wing icing during the war. Post-war, General Electric was converting for the 1950s consumer economy. The Schenectady lab was given two small home freezers the company was thinking about putting on market. Vincent Schaefer, a chemist and meteorologist, was curious to see if they could be used in the lab as home-brew cloud chambers. On a hot day in July, he tried to get one cooler by putting dry ice into it. The cold fog swirling around inside instantly transformed to ice crystals, which fell to the bottom. After playing around, Schaefer found that a single grain of dry ice could, as he wrote in his notebook, “flood the chamber with ice crystals.”

Schaefer and others at the lab (including Bernard Vonnegut, whose inspiration was to use silver iodide as the seed, perhaps also inspiring his brother Kurt’s fictional “Ice-9”) were eager to try out cloud seeding in the wild. But the idea of scientists from the lab monkeying around with the outside weather gave General Electric’s legal department a bad case of cold feet.

In those days, one entity didn’t have to worry much about getting sued: the military. The scientists signed on as consultants to the Army Signal Corps. Between 1947 and 1952, 225 test flights took place over the Berkshires.

The late 1950s and early 1960s, the heydays of big research, saw grandiose-sounding efforts such as Project Skyfire, an attempt to use cloud seeding to reduce lightning strikes on forests in Montana (a study Californians should perhaps dig out of the library); Project Stormfury, an attempt to weaken tropical cyclones; and Project Cirrus, an attempt to squeeze extra water out of warm clouds in New Mexico.

Statistics and the Skeptics: an epistemological dilemma

In the early 1960s, a well-designed, multi-year study of cloud seeding in Israel concluded that cloud seeding in the north of the country increased rainfall by 11 percent. (It had problems in the south, most likely from desert dust). A 3-year program in South Africa in the 1990s confirmed the possibility of hygroscopic, or warm-cloud seeding.

Physically, there are two processes by which cloud seeding works, the “ice process” and the “coalescence” process. California’s orographic (orographic: “relating to mountains”) is the easier ice-process kind, easier because mountain clouds contain supercooled (< 0° Celsius, but still in liquid form) water vapor, and wind and terrain are typically forcing them to rise and cool anyway. For supercooled liquid water (“SLW”) silver iodine crystals act as seed and accelerant, encouraging airborne ice crystals to grow large and heavy enough to fall as precipitation.

Burt Lancaster in The Rainmaker,, 1956.

Yet the skepticism about cloud seeding has always been there, with three main sources: the checkered past of rain-making and rain-makers in folk history; the complex physics of understanding particles swirling around in the atmosphere; and a conundrum best explained by the Central Paradox of Artificial Intelligence.

Fire and rain

If the first meteorologist ever hired by the U.S. Government, in 1843, had been correct, California would have nothing to worry about: its fires would have created their own rain.

James Pollard Espy is today written off as either a mad pyromaniac or a medicine-show barker. But as a scientist, Espy was the first to describe correctly the physics of convective cloud formation. “Espy’s Equation” is still taught today.

Epsy would have been 30 years old in 1815, “The Year Without a Summer.” That was when Mount Tambora in the Dutch East Indies massively erupted, lifting 180 cubic kilometers of tephra into the atmosphere.

J.M.W. Turner, “The Eruption of the Soufriere Mountains in the Island of St. Vincent, 30th April 1812” © University of Liverpool Art Gallery & Collections, The Bridgeman Art Library

The eruption of Tambora put a vast amount of sulfate aerosol into the atmosphere. It’s measurable in ice cores taken in Greenland:

(The 1810 peak is presumably from an unknown eruption somewhere else in the world.)

These aerosols reflect sunlight back into space, increasing Earth’s albedo and decreasing the amount of sunlight absorbed. The “The Year Without a Summer” was an instance of global cooling. The eruption of Mount Pinatubo in the Philippines in June 1991, studied with modern instruments, showed the volcano injected 20 million tons of sulfur dioxide into the stratosphere, causing cooling estimated at 0.3°C over a period of 3 years. This cooling effect is the basis of geoengineering proposals to put reflective particles into the stratosphere.

The visual effects of Tambora’s eruption were not unlike those of the California wildfires. Particles put in the air by Tambora traveled the world, scattering sunlight and producing bright red-orange sunsets in Europe for three years.

J.M.W. Turner, “Sunset,” circa 1830

There was no color photography at the time, but the unusual skies did influence landscape painters. The effect volcanic dust on artists’ color choice is actually measurable, according to a study published in the Journal of Atmospheric Chemistry and Physics in 2014.

Espy, for whatever reasons, was fascinated by the connection of fire and rain. From the classics, Epsy would have been aware of the common observation that heavy rains often follow volcanic eruptions. Pompeii was buried by ash and pumice falling from Mount Vesuvius. Nearby Herculaneum, however, was buried by a mudslide after a freak torrential rain.

In his work for the U.S. government, Espy collected accounts from Native American tribes who intentionally burned prairie and tracts of forest land, a practice inexplicable to Europeans. Espy’s theory was that the tribes were trying to make it rain. He knew that ground heat can cause moist air to rise and, as it ascends, it will cool and may release water as precipitation.

James Pollard Espy

Espy stepping over the line into historical ridicule, wanted to try it out himself. In 1838, he petitioned the United States Senate for permission to burn a large tract of unoccupied forest land, arguing that it if his experiment worked, the rain might keep the Ohio river navigable all the summer. (Espy was turned down, not so much because the Senators thought he was a crackpot, but because they feared his experiment actually might work, giving Espy power to hold Ohio river hostage.) Espy later proposed burning 40 acres of timber every 20 miles in a 600-mile north-south line along the Rocky Mountains, again with the idea of creating rain. A recent proposal for California has an oddly Espy feel about it: it calls for using controlled burns to create a black-and-green checkerboard around the state, the burned parcels serving as firebreaks to keep unintended fires from growing larger.

Particle physics

Epsy was again almost onto something — the key role of particles in the atmosphere. Without dust in the air, we would have no clouds or rain, just a few high-altitude pure ice clouds. Water droplets form around some impurity, typically dust and sometimes salt, known as condensation nuclei.

Water vapor needs careful coaxing to grow into droplets heavy enough to fall as rain. In size, condensation nuclei must be in a Goldilocks zone — not too big and not too small, and varying with wind and temperature. Nor can there be too many condensation nuclei: they compete with each other for available water vapor; too many means not enough of them will grow large enough. By current science, Espy’s fires would have likely put up too many particles to reliably create rain.

All this complexity has made cloud seeding something of a “black box” technology: if you do this, this usually happens. The evidence that it works largely statistical. These statistics are not of the glaringly obvious, overwhelmingly convincing kind. The epistemological dilemma that a weak signal (such as increased precipitation), if it exists, must be teased out of the loud noise of natural variation. Believers see the signal; skeptics see only the vagary of the weather. Ironically, the same form of argument was used for decades by to dismiss new data points on planet warming.

The statistics are convincing enough that several dozen countries, notably China, but including Indonesia, Malaysia, India, Russia, and Thailand, have national programs in weather modification, including precipitation enhancement. But in the U.S., federal funding for weather-modification research dried up in the budget cuts of the 1980s, and never recovered. It is likely Reagan-era politicians were intent on cutting the budget anyway. But the science establishment handed the budget-cutters a potent excuse, one that brings up the Central Paradox of Artificial Intelligence.

What’s in the box

The Central Paradox of Artificial Intelligence is that systems complicated enough to behave intelligently are not simple enough for humans to understand, while systems simple enough to be understandable to humans are not complicated enough to behave intelligently.

If the black box seems to work, but you’re not sure how, do you use it anyway?

If an AI appears to be working perfectly, but humans are incapable of understanding how it works, do we use it? Some of those calling for “understandable” AI would say no.

Various pathways by which water vapor is transformed into various types of cloud particles and precipitation.

The critics of cloud seeding insisted we know everything that was going on inside the box before we use it.

Researchers into the atmosphere would like nothing better than to have a flowchart of all the complex events going on in the atmosphere, and are working towards it. The question is, if the black box seems to work, but you’re not sure how, do you use it anyway?

In 2008, Chinese officials did not want it to rain on the opening ceremonies at the Olympic stadium. They used cloud seeding to try to head off any rain, and make it come down elsewhere. Many found this humorous, and it was. But in fact it didn’t rain on the Birdcage stadium. But can the Chinese “prove” that their seeding intervention made any difference? There lies the epistemological dilemma in a nutshell.

Hope in the West?

In 2017, the University of Wyoming used specially instrumented aircraft to detail — to the satisfaction of almost any science skeptic — the physics of the progression of silver iodide crystals to falling snow.

The physics of warm cloud seeding, of the sort California would need for fire-suppression, are not as well understood. Any any operational program presents major technical challenges, from identifying what clouds are good candidates for seeding, to delivering the seed (probably milled salt crystals) to the right places in a timely manner. A point of optimism is that these challenges are in areas of California strength — radar, computers, aviation.

The real issue raised by the climate crisis revolves around the word “adaptation.” Is there a limit of what humans are willing to endure? Able to endure? People may well prefer to sit back and complain about the weather in the next few years. But if Californians decide they want to try doing something about it, previous generations have left some tools lying around they may wish to dust off.



Will Bates

Will Bates writes about science, technology, and business. His journalism has appeared in the New York Times, the Wall Street Journal, and numerous magazines.