Industries / Frontier science / Cloud seeding
Methods
How seeding is meant to work: adding particles that make ice or large drops form in a cloud that is not raining as much as it could.
7
Milestones achieved, from 1946 to 2017.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1946
Dry ice makes snow
In July 1946 Vincent Schaefer, at the General Electric Research Laboratory in Schenectady, New York, found that dry ice turned a cloud of supercooled droplets in a freezer into ice crystals. Supercooled droplets are cloud water still liquid below 0 °C. On 13 November 1946 he flew from Schenectady with pilot Curt Talbot, dropped 6 pounds (about 2.7 kg) of dry ice into a supercooled layer cloud, and saw streaks of snow fall from it near Mount Greylock, Massachusetts. Twelve days later he seeded a supercooled ground fog and watched it disappear, by his own account.
townofrotterdam.govjournalofweathermodification.scholasticahq.com
Credit
Vincent Schaefer, with Irving Langmuir, General Electric
1946
Silver iodide found
Bernard Vonnegut, a colleague of Schaefer's at General Electric, found that silver iodide particles start ice forming in supercooled cloud, and published the result in the Journal of Applied Physics in July 1947. Silver iodide can be burned into smoke on the ground or in flares, and it is still the main agent for seeding cold clouds, according to a 2024 paper from China's Hebei weather modification center.
Credit
Bernard Vonnegut, General Electric Research Laboratory
1980
Florida test not confirmed
Dynamic seeding puts in heavy doses of silver iodide so the heat released by freezing makes cumulus clouds grow taller and rain more. The Florida Area Cumulus Experiment of 1970 to 1976 suggested rain increases over a 13,000 km² target in south Florida. A second, confirmatory phase in the summers of 1978 to 1980, with its analyses fixed in advance, did not confirm them.
Lessonan exploratory result needs a confirmatory experiment, specified in advance, before it is relied on.
Credit
William Woodley, Joanne Simpson and colleagues, NOAA
1983
Hurricane seeding dropped
Project Stormfury, run by the US Weather Bureau, the Navy and the National Science Foundation from 1962 to 1983, seeded clouds just outside a hurricane's eyewall with silver iodide to weaken its strongest winds. Seeding was tried in four hurricanes on eight days, the last in Ginger in 1971, and winds fell 10 to 30 percent on four of those days. Observations in the 1980s showed that hurricanes hold too little supercooled water, and that unseeded storms form new outer eyewalls on their own, so the program was ended.
Lessona seeding result means nothing until it can be told apart from what the storm would have done anyway.
Credit
Project Stormfury, US Weather Bureau and NOAA
1984
Propane clears an Olympic airport
At the 1984 Winter Olympics in Sarajevo, ground dispensers sprayed liquid propane into fog colder than 0 °C. Evaporating propane chills the air enough to make ice crystals, which grow and fall out. Of 11 seedings at the airport, the team counted nine successes; on 6 and 7 February 1984 the cleared fog let 30 aircraft land and take off before the opening ceremony.
Credit
D. Milosevic and colleagues, Republic Hydrometeorological Institute of Serbia, with N. Fukuta, University of Utah
1990
Salt flares for warm clouds
Hygroscopic seeding adds salt particles that attract water, so a cloud makes some larger drops that collide and grow into rain. In South Africa a research jet found unusually large drops in a storm growing over a Kraft paper mill; that led to hygroscopic flares, field trials from October 1990, and a randomized experiment that showed more radar-measured rain from seeded storms. The UAE's National Center of Meteorology still calls hygroscopic flares its traditional option.
Credit
Graeme Mather (CloudQuest, Nelspruit) and colleagues at South Africa's Weather Bureau
2017
Airports retire fog clearing
Medford, Oregon, lifts crushed dry ice under a tethered helium balloon to clear freezing fog, and Missoula, Montana, sprays liquid carbon dioxide from trucks about ten times a winter. Anchorage, Salt Lake City, Boise and Reno, which once fought winter fog, no longer do, because newer jets with precision landing aids can operate in very low visibility.
Lessona better landing system can remove the need for a weather fix.
Credit
Rogue Valley International-Medford Airport (CASPER balloon, first flown 2010); Missoula airport
Watch for
Nano-salt seeding powder
The UAE's National Center of Meteorology says a nanomaterial powder developed through its rain enhancement research program is up to three times as effective as hygroscopic flares, and disperses as soon as an igniter opens its cap. The material, sodium chloride with a titanium dioxide shell, comes from Linda Zou's group at Khalifa University, which has filed two patents. In a February 2026 preprint, Michigan Tech's Pi Chamber injected it and plain milled salt into a laboratory cloud and saw drops up to 60 micrometres form.
Achieved whena field trial comparing seeded and unseeded clouds, evaluated by a group outside the center and Khalifa University, is published.
Claimed by
UAE National Center of Meteorology, July 2026
Open
01
A world seedable-cloud atlas
Monthly maps, for each region where seeding programs run, of how often clouds hold supercooled liquid water at the temperatures silver iodide works (about −4 to −20 °C), built from satellite observations rather than model output, and covering programs outside the US West.
The briefCredit · unclaimed02
Count clearable fog hours
An airport-by-airport count of the hours each winter with fog colder than 0 °C, visibility below landing limits and no falling snow, which are the conditions in which propane, dry ice or liquid carbon dioxide can clear fog, for every airport with 20 or more years of public weather reports.
The briefCredit · unclaimed
Open 01 · the brief
A world seedable-cloud atlas
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
Monthly maps, for each region where seeding programs run, of how often clouds hold supercooled liquid water at the temperatures silver iodide works (about −4 to −20 °C), built from satellite observations rather than model output, and covering programs outside the US West.
Why it is still open
Seedability maps exist one region at a time, each by its own method. For the interior western United States there are two built from regional climate models, not observations: Mazzetti, Geerts, Xue, Tessendorf and colleagues (2021) found seedable conditions in 20 to 30 percent of cold-season hours on many ranges, and Adhikari, Geerts, Tessendorf and Xue (2025) mapped supercooled water below −6 °C over 43 years on a 4 km grid. For Saudi Arabia, a 2026 preprint from the national seeding program picks seedable clouds from Meteosat satellite data and matched 82.9% of more than 10,000 operational seeding events from 2022 to 2025. A global CloudSat and CALIPSO climatology of mixed-phase clouds published in September 2026 sorts clouds by type, not by seeding temperature or target area.
ckanprod.data-commons.k8s.ucar.edudoi.orgdoi.orgacp.copernicus.org
Built with what exists
CloudSat's 2B-CLDCLASS-LIDAR cloud type and phase product and its ECMWF-AUX temperature profiles (from June 2006), Python with xarray, and the published Mazzetti and Adhikari seedability values for comparison.
First step
Pull 2B-CLDCLASS-LIDAR profiles for a box around Idaho's Payette basin for 2006 to 2010 and count supercooled liquid layers between −4 and −20 °C by month.
Done looks like
Monthly seedable-cloud frequencies for at least 30 named seeding target areas on at least four continents, by one method, published with code and data under an open licence, with the difference from the model values stated for at least 10 US mountain ranges and from the Saudi satellite method for the Saudi target areas.
Open 02 · the brief
Count clearable fog hours
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An airport-by-airport count of the hours each winter with fog colder than 0 °C, visibility below landing limits and no falling snow, which are the conditions in which propane, dry ice or liquid carbon dioxide can clear fog, for every airport with 20 or more years of public weather reports.
Why it is still open
Airports have split on fog clearing, with Medford and Missoula still doing it while Anchorage, Salt Lake City, Boise and Reno stopped, and no airport-by-airport count of clearable fog was found. Raw counts are unreliable: a July 2025 study by Scott Landolt of NCAR found US automated stations report freezing fog during snow when it is often not present.
Built with what exists
The Iowa Environmental Mesonet archive of airport weather reports from around the world, NOAA's one-minute ASOS station data for a misreport filter of the kind Landolt tested, and Python.
First step
Download 20 winters of reports for Medford (KMFR) and Missoula (KMSO) and count the hours that meet the fog-clearing conditions.
Done looks like
A public table for at least 200 airports covering the winters 2005 to 2025, with the snow-misreport filter checked against human-observed Canadian reports, and Medford and Missoula ranked against the rest.