Industries / Frontier science / Cloud seeding
Public concerns
The fears people raise about seeding, from floods and stolen rain to silver in the soil, and how inquiries, scientists and courts have answered them.
7
Milestones achieved, from 1916 to 2025.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1916
Rainmaker meets a flood
In December 1915 San Diego's city council made a verbal "no rain, no pay" deal to pay Charles Hatfield $10,000 to fill Morena Reservoir. Nearly 30 inches of rain fell in January 1916, the region's wettest period on record, and the Lower Otay Dam failed on 27 January; the county coroner estimated 50 deaths, a figure later revised to about 20. The city refused to pay unless Hatfield took liability for the damage, a judge called the flood "an act of God" in damage suits against the city, and Hatfield's own suit was dismissed in 1938.
LessonClaiming credit for rain brings the bill for the flood.
Credit
Charles Hatfield; San Diego city council
1972
Rapid City flood inquiry
On 9 June 1972 the South Dakota School of Mines flew two seeding flights near the Black Hills for a Bureau of Reclamation experiment; that night a flash flood killed 238 people in Rapid City and nearby. A state Board of Inquiry found that "in the absence of cloud seeding, the result would have been the same". Heirs of five victims sued the United States in 1975; the suit was dismissed in 1982 on federal immunity, so no court ruled on the cause. A poll in late 1972 found about 30% of South Dakotans saw some link between seeding and the flood.
LessonAn inquiry can clear seeding on the science and still leave much of the public unconvinced.
Credit
Board of Inquiry (Pierre St. Amand, Ray Jay Davis, Robert Elliott); account by Arnett Dennis of the Institute of Atmospheric Sciences
1993
Montana answers "cloud rustling"
Eastern Montanans worried that seeding in western North Dakota was "cloud rustling" that took rain meant for them. In 1993 Montana's legislature passed Senate Bill 72, which makes a seeding permit applicant file an environmental impact statement, hold a public meeting, and show $10 million in bonds, insurance or cash to cover unexpected damage.
Credit
Montana Legislature
2017
Silver stays near background
Silver iodide is the usual agent for winter snow seeding, and people ask whether its silver builds up. Snowy Hydro's monitoring at 13 Australian generator sites from 2004 to 2009 found mean silver and indium in soil always well below guideline values. A Boise State University study of more than 4,000 Idaho snow samples found about 2.8 parts per billion of silver within 5 m of ground generators, no seeding trace beyond 80 km, and concentrations four orders of magnitude below US EPA secondary standards. A 2016 laboratory study at Complutense University of Madrid saw moderate harm to soil and water organisms only at its highest test dose, 12.5 micromolar.
LessonGAO judged in December 2024 that the evidence, "limited to a handful of recent studies", shows no concern at current levels, but that the effect of wider use is not known.
journalofweathermodification.scholasticahq.comscholarworks.boisestate.edupubmed.ncbi.nlm.nih.govtdlr.texas.gov
Credit
Cardno Ecology Lab and Snowy Hydro; James Fisher, Boise State University, with Idaho Power; Fajardo and colleagues, Madrid
2018
Iran alleges cloud theft
On 2 July 2018 Gholamreza Jalali, head of Iran's Passive Defense Organization, said Israel and a neighbouring country were making clouds over Iran barren and "stealing" clouds and snow. Ahmad Vazifeh of Iran's Meteorological Organization replied that it was not possible for a country to steal snow or clouds, and pointed to years of drought.
Credit
Iran Meteorological Organization
2024
Dubai floods blamed on seeding
On 14 and 15 April 2024 record rain hit the United Arab Emirates and northern Oman, killing at least 20 people in Oman and 4 in the UAE, and speculation spread on social media that the UAE's seeding program caused it. A rapid study by World Weather Attribution found that seeding cannot add the moisture that made the storm unusual and judged that it had no significant influence.
LessonReports disagreed on whether seeding flights flew in the days before, so the study answered on the physics rather than the flight record.
Credit
World Weather Attribution, 25 April 2024
2025
Texas flood and Rainmaker
Flash floods on 4 July 2025 in the Texas Hill Country killed more than 100 people, and social media posts blamed Rainmaker Technology, a company seeding in south-central Texas. Its chief executive, Augustus Doricko, said its last flight was on 2 July and the two clouds it seeded broke up that afternoon; meteorologists tied the rain to the remains of Tropical Storm Barry, and PolitiFact rated the claim false.
LessonNOAA posts reports about two months after each fiscal quarter ends, so in the first days the public had only the company's word on where it had flown.
Credit
PolitiFact, 7 July 2025
Watch for
Seeding agents and bee learning
Laboratory tests found that silver iodide and potassium iodide given to honey bees at low and high doses slowed learning in shock-avoidance and odour tests and changed return times in a flower-patch test.
Achieved whenanother laboratory repeats the effect at doses matched to silver measured in seeded areas, or a field study finds it near seeding sites.
Claimed by
Wincheski and colleagues, Oklahoma State University and the University of Tulsa, in the journal Insects, November 2025
Citizen reports of seeding
Florida's 2025 law lets anyone report observed weather modification online to the state environment department, which must screen the reports and investigate those that warrant it. SkyLedger, a volunteer website, tags aircraft owned by known seeding operators on a live map and lets users flag aircraft as evidence.
Achieved whena citizen report leads a regulator to confirm seeding that had not been reported.
Claimed by
Florida Legislature, June 2025; SkyLedger, 2026
Open
01
A "was it seeded?" lookup
A public website that answers, for any US place and date, whether official records show cloud seeding nearby in the days before, links the source reports, and says plainly when reports for that period are not yet filed.
The briefCredit · unclaimed02
An open silver ledger
One open dataset of every published measurement of silver, and of indium tracer, in snow, rain, soil, sediment, water and living things near seeding programs, with distance from the source, method, detection limit and local background, plus laboratory toxicity results with their doses in common units.
The briefCredit · unclaimed
Open 01 · the brief
A "was it seeded?" lookup
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public website that answers, for any US place and date, whether official records show cloud seeding nearby in the days before, links the source reports, and says plainly when reports for that period are not yet filed.
Why it is still open
NOAA's archive cannot be searched by place and posts reports about two months after each fiscal quarter, and GAO found it incomplete. Donohue and Lamb's dataset geocodes each project's target area to a single point for a map figure, not to an area a query can test. SkyLedger maps live aircraft and user flags, not official records.
Built with what exists
Donohue and Lamb's dataset; QGIS or PostGIS to trace each report's target-area map into a polygon; TDLR's map of Texas project areas; an AI model reading scanned PDFs with a person checking (see the Note "Seven more flagged"); the OpenSky Network's archive of aircraft position reports, used only to cross-check the flight dates operators report.
First step
Trace the five Texas project areas from TDLR's map and the NOAA reports, then answer the 2 to 4 July 2025 Kerr County question from records alone.
Done looks like
Answers for every state for 2000 to 2025; a written test on 10 publicised flood claims since 2000, each showing what the records said and when they became public; no more than 5 wrong answers in 100 random place-and-date checks against the source reports.
Open 02 · the brief
An open silver ledger
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
One open dataset of every published measurement of silver, and of indium tracer, in snow, rain, soil, sediment, water and living things near seeding programs, with distance from the source, method, detection limit and local background, plus laboratory toxicity results with their doses in common units.
Why it is still open
GAO called the field evidence "a handful of recent studies". Summaries exist, but they quote conclusions, not values: Idaho's Water Resource Board, in a February 2024 fact sheet, says more than 20 studies and data reviews found no evidence of harm. The studies themselves sit in separate reports with different units: parts per billion in Idaho snow, guideline ratios in Australian soil, up to 140 micrograms per square metre of silver iodide estimated from hail suppression in Serbia, and micromolar doses in laboratory tests from Madrid (2016) and Oklahoma (2025).
tdlr.texas.govidwr.idaho.govscholarworks.boisestate.edupubmed.ncbi.nlm.nih.govpubmed.ncbi.nlm.nih.gov
Built with what exists
Crossref and PubMed searches; an AI model reading scanned PDFs with a person checking (see the Note "Seven more flagged") to pull values from tables; Python and pandas for unit conversion; Zenodo for the release.
First step
Code the values from the four studies above and the 2009 Snowy Mountains toxicity assessment in the Journal of Weather Modification, and publish the schema.
Done looks like
At least 30 studies coded, each value traced to a page and table; one chart that sets field concentrations beside the doses that caused harm in laboratory tests; the dataset cited or reviewed by a state seeding program or an environmental agency.