Is cloud seeding harmful? Silver iodide, flood claims and what seeding can really do

Mostly, no. Cloud seeding adds small amounts of silver iodide, salt or dry ice to clouds that already hold water, to coax a little more rain or snow out of them. Monitoring around programmes that have run for decades has generally found silver in snow, soil and streams at levels close to natural background and well below those regarded as a risk to people or wildlife. Seeding cannot make clouds where there are none, and the claims that it caused the floods in Dubai in April 2024 or in central Texas in July 2025 do not survive contact with the meteorology.

That is not the end of the matter. The harder questions are whether seeding works well enough to be worth paying for, whether one valley's extra snow comes at the expense of the next, and who gets a say before the aircraft take off. They are close relatives of the questions now asked about far larger schemes to adjust the climate.

What is cloud seeding?

Many clouds hold droplets of liquid water at temperatures well below freezing. Pure water struggles to freeze without a speck of something to freeze onto, so these "supercooled" droplets can drift over a mountain range and give up less snow than they might. Cloud seeding supplies the specks.

The idea came out of a General Electric laboratory in Schenectady, New York. In 1946 Vincent Schaefer found that dry ice dropped into a chamber of supercooled cloud filled it with ice crystals, and in November that year he scattered dry ice from an aircraft over western Massachusetts and watched snow fall. His colleague Bernard Vonnegut soon showed that silver iodide, whose crystal structure is strikingly close to that of ice, worked as a seed too and could be burned into a fine smoke.

Most seeding today belongs to one of two families:

• Glaciogenic seeding uses silver iodide, occasionally dry ice, to turn supercooled droplets into ice crystals that grow and fall as snow. It is the standard method for winter clouds over mountains, and for hail suppression.
• Hygroscopic seeding uses fine salt particles in warmer clouds, encouraging bigger droplets that merge and fall as rain. It is the main approach in hot, dry countries such as the United Arab Emirates.

The seeds arrive by flares on racks under an aircraft's wings, by ground generators on ridges that burn a silver iodide solution into rising air and, in China, by rockets and artillery shells. Drones are a newer addition.

Who uses it, and why

In the United States the biggest customers are water managers in the West. Utah, Colorado, Wyoming, Idaho and Nevada all run winter programmes to thicken the snowpack that feeds rivers in spring, and water agencies further down the Colorado River have helped pay for seeding upstream. Texas and North Dakota have long-running programmes for rain and hail suppression, funded mostly through counties and water districts. In Canada, insurers have paid for hail suppression around Calgary since the 1990s.

China runs by far the largest operation: in 2020 its government announced plans to extend weather modification over more than five million square kilometres by 2025, over half the country. The UAE has seeded clouds for more than two decades and funds an international research programme on rain enhancement.

Does cloud seeding work?

Every cloud happens once, so nobody can watch the same cloud with and without seeding. Rain and snow vary hugely from storm to storm, which buries a modest effect in noise. Operators often claimed gains of around 10 per cent; independent scientists found the evidence thin. In 2003 a US National Research Council report concluded that there was still no convincing scientific proof that deliberate weather modification worked, and called for better research.

Two projects shifted the picture for winter clouds over mountains. The Wyoming Weather Modification Pilot Project ran a randomised experiment from 2005 to 2014; its analysts concluded that seeding suitable storms probably added a modest amount of snow, though the statistics could not rule out chance. Then, in early 2017, the SNOWIE experiment in Idaho flew seeding aircraft in zigzags across the wind while radar watched, and lines of snow appeared downwind in the same zigzag pattern. Papers in 2018 and 2020 traced the chain from silver iodide to ice crystals to snow on the ground.

The broad position now, roughly the one the World Meteorological Organization takes, is that seeding can squeeze some extra snow from the right kind of winter cloud over high ground: a modest addition, not a transformation. Evidence for salt seeding of warm clouds and for seeding summer storms is weaker, and hail suppression results are mixed. Nobody serious claims seeding can end a drought, because a drought is a shortage of the clouds you would need to seed.

Is cloud seeding harmful? The worries one by one

Silver in soil and water

Dissolved silver is toxic to fish and other aquatic life at low concentrations, so the question is a fair one. Silver iodide, however, barely dissolves in water, and seeding spreads a small quantity of it thinly across a large catchment. Monitoring around long-running programmes in the American West and in Australia's Snowy Mountains has generally found silver at or near the limits of detection, often hard to tell apart from natural background. Scientific assessments have judged the ecological risk to be low, while urging continued monitoring where silver could build up in sediments.

Human health

For the public, the realistic exposure is trace amounts in rain or snow, far below the limits set for drinking water. Those who make and handle the flares face an ordinary occupational safety question. Salt and dry ice raise no health concerns at the quantities used. For specific worries, a national environmental or health regulator is a better guide than social media.

Can cloud seeding cause floods?

Not on the scale people have claimed. A large storm carries an immense quantity of water. Seeding works at the margins of individual clouds and, at best, tips a little more of their existing water out as rain or snow; it adds no moisture to the atmosphere. Responsible operators also stand down when flood or avalanche warnings are in force, and many permits require it.

Dubai, April 2024. On 16 April 2024 the UAE recorded its heaviest rainfall since records began in 1949, with some places receiving more than a typical year's rain in a day. Dubai's airport flooded and posts online blamed cloud seeding, though the national meteorology centre said none took place during the storm. More to the point, global models had forecast the system days ahead, and the same weather had already caused deadly flooding in neighbouring Oman. A rapid study by the World Weather Attribution group concluded that downpours like it have probably become heavier in the region as the climate warms.

Central Texas, July 2025. After the flash floods on the Guadalupe River in early July 2025, which killed more than a hundred people, attention turned to a seeding company that had flown a small operation in the region a couple of days before. Meteorologists pointed out that the rain came from a deep plume of tropical moisture and that the totals were far beyond anything seeding could produce. The company denied any link. The episode still fed calls for bans in Texas and other states.

Stealing the neighbours' rain

The oldest objection is that seeding one area robs another downwind. The physics makes this unlikely on any large scale: clouds turn only a fraction of the water vapour passing through them into precipitation, and seeding nudges that fraction slightly. Studies that looked for downwind shortfalls have mostly not found them. That has not stopped officials in more than one dry country from accusing neighbours of stealing their rain. When water is scarce, a plausible villain is politically convenient.

Who regulates cloud seeding?

Regulation is patchy, and much of it is older than today's arguments.

• United States, federal. Under the Weather Modification Reporting Act of 1972, anyone seeding clouds in the US must report the project to NOAA before it starts and afterwards. NOAA does not approve or refuse projects; it collects the paperwork, and the records are public.
• United States, states. Real control sits with the states. Several, including Texas, Utah, Colorado and North Dakota, require licences or permits. Tennessee in 2024 and Florida in 2025 went the other way and banned deliberate weather modification outright.
• International. The ENMOD Convention, adopted at the UN in 1976, bans hostile military use of environmental modification. It followed Operation Popeye, in which the US seeded clouds over the Ho Chi Minh Trail from 1967 to 1972 to turn supply routes to mud. Peaceful seeding has no binding international regime; the WMO offers guidance.

Disputes over liability are older than any statute. In 1950 New York City, short of water, hired a meteorologist to seed clouds over its Catskills reservoirs; local resort owners sued, saying the rain was ruining their summer trade. The judge let the programme continue, putting the city's need for water first.

How is it different from solar geoengineering?

Cloud seeding is often lumped in with geoengineering, and some of the new state laws treat them as one thing. They differ hugely in scale and purpose.

Seeding acts on individual clouds for hours and has no meaningful effect on the planet's temperature. Solar geoengineering aims to cool the whole Earth, with effects that would last only as long as someone kept it going. The closer parallel may be ocean iron fertilisation, another attempt to seed a natural system for a benefit, which came with bold claims, thin evidence and a governance fix that arrived after the controversy rather than before it.

What they share is a problem of consent. A water district that pays for seeding and a farmer downwind who believes it took her rain have different stakes and very different power. Elizabeth Kolbert's Under a White Sky (Amazon (book)) is a good account of how human fixes for nature tend to call for further fixes.

Two more questions people ask

Is cloud seeding the same as "chemtrails"?

No. The white lines behind airliners are contrails, thin ice clouds formed from water vapour in jet exhaust, and the theory that they are secret chemical spraying has no evidence behind it. Cloud seeding is real, and in the US it is reported to NOAA, with the records open to anyone.

Can cloud seeding stop hurricanes?

The US tried with Project Stormfury between 1962 and 1983 and gave up. Hurricanes turned out to contain too little supercooled water for seeding to make a difference.

The question worth asking

The useful test for any seeding programme is the one worth applying to bigger climate schemes: who paid for it, who was told in advance, and who could have said no? On those questions, cloud seeding's record is better than its reputation online, and rather worse than its promoters like to admit.