Cloud seeding is not only still done, it is more widespread and better funded than at any point in its seven-decade history. Dozens of countries actively run operational programs, from the western United States boosting mountain snowpack to the United Arab Emirates trying to squeeze rain from desert skies. What has changed is not the ambition but the evidence base: recent field campaigns have, for the first time, directly measured the snow and rain that seeding produces, moving the practice from “probably works, hard to prove” to something closer to quantified results.
Where Cloud Seeding Happens Today
In the United States, winter cloud seeding programs operate across much of the Mountain West. Wyoming’s Water Development Commission funds an ongoing program over the Medicine Bow and Sierra Madre ranges, using aircraft to seed winter storms and boost snowpack that feeds river systems downstream. Idaho, Utah, Colorado, and California all run or fund similar efforts aimed at padding reservoir levels. In the Great Plains, Kansas has maintained a summer cloud seeding program focused on hail suppression and modest rainfall increases for agriculture.
Internationally, the scale is even larger. The UAE’s national rain-enhancement program is one of the most heavily invested in the world, routinely seeding convective clouds with hygroscopic materials. China operates what is likely the biggest weather modification program on Earth, deploying thousands of ground-based generators and aircraft across multiple provinces. Australia, Thailand, India, France, Spain, Serbia, and several African nations all run programs of varying size. Alberta, Canada, has operated a hail-suppression program since the 1990s, seeding thunderstorms from aircraft to protect crops and property.
How Modern Cloud Seeding Works
The basic idea is straightforward: you introduce tiny particles into clouds that help water droplets or ice crystals form more efficiently than they would on their own. The details depend on what kind of cloud you are seeding.
For cold clouds, the standard agent is silver iodide. Its crystal structure resembles ice closely enough that supercooled water droplets freeze onto it. Silver iodide particles trigger ice formation through several pathways depending on temperature, humidity, and how the particles encounter cloud droplets. Laboratory work at Colorado State University showed that the two most common operational silver iodide formulations can initiate freezing through contact with droplets, through condensation and subsequent freezing, or through direct deposition of water vapor as ice. Which pathway dominates depends on conditions inside the cloud, particularly how much moisture is available.
For warm clouds in tropical or arid regions, the approach is different. Operators release hygroscopic (moisture-attracting) particles, often based on sodium chloride or calcium chloride flares, into the base of growing cumulus clouds. These large salt particles serve two purposes: they can grow quickly into raindrop-sized embryos, and they also suppress the activation of smaller natural particles, which paradoxically helps the remaining droplets grow bigger and fall as rain faster. Numerical modeling has confirmed both of these mechanisms, sometimes called the “raindrop embryo effect” and the “competition effect.”
The Evidence That It Actually Produces Precipitation
For decades, the central frustration with cloud seeding was the difficulty of proving it worked. Clouds are chaotic, storms vary enormously, and you can never know exactly what would have happened without seeding. That situation changed substantially with the SNOWIE project (Seeded and Natural Orographic Wintertime Clouds: the Idaho Experiment), which ran field campaigns in Idaho’s Payette Basin starting in 2017.
SNOWIE used ground-based radar alongside precipitation gauges to track what happened after aircraft released silver iodide into winter mountain clouds. Researchers were able to isolate areas of precipitation that could be unambiguously attributed to seeding, separating them from natural snowfall. For the cases they analyzed, individual seeding runs lasting 20 to 86 minutes generated between roughly 100 and 275 acre-feet of water each. Precipitation gauges on the ground recorded increases between 0.05 and 0.3 millimeters as seeded precipitation bands passed overhead. These are modest amounts per event, but they represented a breakthrough: the first time anyone had directly traced snowfall from seeding agent release to ground accumulation using physical measurements rather than statistical inference.
A follow-up case study from the same SNOWIE dataset looked more closely at the microphysical changes inside seeded versus unseeded clouds. Seeded clouds showed markedly higher ice water content and concentrations of large ice particles, while liquid water content and cloud droplet numbers dropped by about 30 percent, indicating that the silver iodide was efficiently converting liquid water to ice. The authors concluded this confirmed the basic glaciogenic seeding mechanism was working as intended.
How Much Extra Water Can You Get?
The per-event numbers from SNOWIE are small, but cloud seeding programs run dozens of operations per winter season, and the cumulative effect adds up. Modeling work simulating a full winter season (2019–2020) of aerial seeding over ranges in southern Wyoming and northern Colorado estimated that the program added over 10 millimeters of liquid-equivalent precipitation at the highest-elevation sites. Across the North Platte and Little Snake River basins combined, the seasonal total came to roughly 9,500 acre-feet of additional water, with an ensemble range spanning about 6,000 to 14,000 acre-feet depending on modeling assumptions. For context, an acre-foot is about 326,000 gallons, enough to supply roughly two average American households for a year. So a single season’s seeding in one corner of Wyoming generated water equivalent to a small-town supply.
These numbers do not make cloud seeding a silver bullet for drought. The increases typically run in the range of 5 to 15 percent above what natural storms would deliver, and they depend entirely on having the right clouds to seed. You cannot create precipitation from a clear sky. But for water managers watching reservoir levels and snowpack forecasts, a consistent 5 to 15 percent bump over a winter season is a meaningful addition, particularly given that the cost per acre-foot of water from seeding is low compared to alternatives like desalination or new reservoir construction.
Cloud Seeding in Arid Climates
The UAE and other arid-region programs face a tougher challenge: fewer seedable clouds and less atmospheric moisture to work with. Interestingly, modeling research suggests that some newer seeding agents may actually be more effective in dry conditions than in humid ones. A study comparing cloud seeding in three climate zones (the UAE’s arid climate, Thailand’s maritime climate, and Serbia’s continental climate) found that a sodium chloride/titanium dioxide nanostructure agent produced the greatest precipitation increase in the arid case, generating about 1.4 times more surface rainfall than conventional sodium chloride seeding.
This is counterintuitive at first. You might expect seeding to work best where there is plenty of moisture. But in arid convective clouds, the atmosphere is often so short of natural condensation nuclei that introduced particles face less competition and can have an outsized effect on droplet growth. The catch is that the absolute amounts of rainfall remain modest, since there is less total moisture available. Still, in a country like the UAE where every additional millimeter of rain matters for groundwater recharge, even small gains are worth pursuing.
Hail Suppression Programs
Boosting precipitation is only half the story. A significant share of operational cloud seeding worldwide aims not to make more rain but to reduce hail damage. The logic is that by seeding thunderstorms with silver iodide, you create many small ice particles that compete for the available supercooled water. Instead of a few hailstones growing large enough to cause serious damage, you get many smaller ones that melt before reaching the ground or cause much less harm.
Alberta’s long-running hail-suppression program has been analyzed using a decade of radar data. When seeded and unseeded portions of storms were compared, about 60 percent of seeded cases showed lower peak hail intensity indicators than their unseeded counterparts. Around 8 to 20 percent showed no change, and 17 to 30 percent actually showed higher values, suggesting seeding does not work on every storm. The statistical differences between seeded and unseeded groups were significant overall, though, and the effect became more pronounced about 30 minutes after seeding began.
In Europe, ground-based silver iodide generator networks have been used for hail suppression in France and Spain for decades. Analysis of those programs suggests that when generators are activated three hours before hail reaches the ground, using a network spaced about 10 kilometers apart, the energy of hailfall on the most damaging days drops by roughly half. That is a substantial reduction in crop and property losses, which explains why farming regions continue to fund these programs even when the science around precipitation enhancement remains debated.
Does Cloud Seeding Steal Rain from Downwind Areas?
One of the most common concerns about cloud seeding is whether it robs downstream communities of their rainfall. If you wring extra water out of a cloud over one mountain range, does the area on the other side get less? This worry has been a source of legal disputes and public opposition since the earliest days of weather modification.
The available evidence is reassuring on this point, though not as extensive as anyone would like. An analysis of Kansas’s cloud seeding program found an average rainfall increase of about 7 percent in targeted regions but no significant reduction in downwind areas, suggesting that seeding was not producing a detectable “rain steal” effect. The study noted this was consistent with earlier findings from Alberta’s program, which also reported no measurable downwind decreases.
The physical reasoning behind this is that most cloud seeding programs target only a fraction of the available moisture in any given storm system. Clouds passing over a seeded area continue to develop and produce precipitation as they move downwind. In mountain settings, orographic lift keeps forcing moist air upward, regenerating cloud water after seeded precipitation falls. That said, the question becomes harder to answer definitively in regions where cloud seeding programs are very large or where multiple programs operate in sequence along the same storm track. Most atmospheric scientists consider the downwind-depletion risk low for current program scales, but it is not zero, and monitoring it matters as programs expand.
Environmental and Health Concerns Around Silver Iodide
Silver iodide is the workhorse of cold-cloud seeding, and the environmental question people most often raise is whether dispersing it into the atmosphere contaminates soil and water. The amounts involved per seeding operation are small, typically a few grams of silver iodide burned per hour per generator, but programs that operate for decades in the same watersheds do accumulate detectable silver in soil and surface water.
Field measurements near a seeding generator site in Steamboat Springs, Colorado, found that deposited silver tends to stay in the top two centimeters of soil. At high concentrations near the generator, researchers noted higher organic matter levels, moisture content, and microbial populations, along with faster soil respiration. The interpretation was that silver at those concentrations was inhibiting the breakdown of organic matter, which caused it to accumulate. Importantly, the study concluded that the silver levels actually found in seeded target areas (as opposed to right next to a generator) should not cause overt effects on the soil environment.
Laboratory ecotoxicology work paints a slightly more cautious picture. Testing silver iodide on freshwater and soil organisms showed moderate adverse effects at the highest concentrations tested, but even at much lower levels matching the reference concentration used in environmental monitoring (0.43 micromolar), silver iodide significantly reduced photosynthetic activity in cyanobacteria and green algae, with respiration dropping by about 80 percent and net photosynthesis by about 40 percent in both phytoplankton strains tested. Soil bacteria showed a moderate decrease in cell viability, though a common soil worm showed no effects on growth or survival. The study’s authors noted that these impacts could become meaningful if seeding is repeatedly applied in a specific area and large amounts of material accumulate over time.
In practice, operational programs disperse silver iodide over wide areas and at concentrations far below the thresholds that caused problems in the lab. The silver concentrations measured in rivers and lakes downwind of long-running programs have generally been at or below detection limits. But the laboratory findings are a reminder that this is not an entirely benign process, and that environmental monitoring should be a standard part of any program running for years in the same watershed.
Why Public Perception Lags Behind the Science
Cloud seeding occupies an unusual space in public awareness. Many people assume it was an experimental curiosity of the mid-twentieth century that never panned out. Others have encountered it only through conspiracy theories linking it to chemtrails or covert weather warfare. Neither picture is accurate, but both persist partly because the field spent decades unable to produce the kind of clean, controlled evidence that settles questions in other areas of science.
The difficulty is inherent to the problem. You cannot put the atmosphere in a laboratory. Every cloud is different, natural variability is enormous, and the “control” case (what would have happened without seeding) is always hypothetical. Early randomized experiments in the 1960s and 1970s produced mixed results, and some high-profile projects were later criticized for statistical shortcomings. This left a lingering reputation of ambiguity that the field is only now shaking off through campaigns like SNOWIE, which used modern radar and in-situ measurements to physically trace seeded precipitation rather than relying solely on statistical comparisons between seeded and unseeded periods.
The other factor is that cloud seeding programs are typically run by state water agencies or private contractors, not headline-grabbing research institutions. They operate quietly, season after season, and rarely generate the kind of media attention that would update public understanding. Meanwhile, the conspiracy-theory version of cloud seeding spreads readily on social media, often conflating operational weather modification with unrelated geoengineering proposals like stratospheric aerosol injection. The two are completely different in scale, mechanism, and intent, but their conflation muddies the conversation.
Newer Seeding Agents and Techniques
Silver iodide has been the dominant seeding agent since the late 1940s, but it is not the only option, and researchers are actively exploring alternatives. Hygroscopic flares containing salts like sodium chloride and potassium chloride are widely used for warm-cloud seeding in tropical programs. More recently, engineered nanostructures combining sodium chloride cores with titanium dioxide shells have shown promise in modeling studies, particularly for arid environments where conventional salt particles underperform. In the three-climate-zone study mentioned earlier, this core/shell nanostructure outperformed plain sodium chloride in the arid UAE test case by a factor of 1.4 in terms of accumulated surface precipitation.
Drone-based seeding is another area of active development. Traditional programs rely on piloted aircraft or ground-based generators, both of which have limitations: aircraft are expensive to operate in bad weather, and ground generators cannot target specific cloud layers precisely. Drones could potentially deliver seeding agents to exact altitudes and locations within a cloud, improving efficiency. Several countries, including the UAE and China, have tested drone-based delivery, though the technology is still maturing and faces challenges with payload capacity and flight duration in turbulent conditions.
The computational side has advanced as well. Modern programs increasingly use high-resolution weather models to decide when, where, and whether to seed. The Wyoming modeling work, for instance, used a weather-modification-adapted version of a widely used forecasting model to simulate an entire season of operations and estimate the water yield. This kind of modeling lets program managers make more targeted decisions about which storms are worth seeding and helps them estimate the return on investment for each operation, moving the practice from art toward engineering.
What Cloud Seeding Cannot Do
Perhaps the most important thing to understand about cloud seeding is what falls outside its capabilities. It cannot create rain from a cloudless sky. It cannot end a drought if the atmospheric pattern is not producing seedable clouds. It cannot meaningfully change regional climate. And it cannot redirect storms or steer weather systems. All it can do is nudge existing clouds to produce somewhat more precipitation, or somewhat smaller hail, than they would have naturally.
This means cloud seeding is a supplemental water-management tool, not a replacement for conservation, infrastructure investment, or emissions reduction. In a good year with frequent storms, a seeding program might add a useful boost to snowpack. In a year dominated by high-pressure ridges that block storm systems entirely, the same program sits idle because there are no clouds to work with. The places where cloud seeding makes the most consistent difference tend to be mountain regions with reliable orographic cloud formation in winter, where seeding can systematically enhance snowfall across many small events over the season. In flat terrain with sporadic summer convection, the opportunities are fewer and the results are harder to quantify.
For anyone living in a region with an active cloud seeding program, the practical takeaway is straightforward: the program is almost certainly producing some additional water, the environmental risks at current scales appear low but are not zero, and the rain falling on your neighbor’s field was not stolen from yours. The technology is real, it is expanding, and the evidence supporting it is stronger than it has ever been, even if it remains a modest tool for a problem that demands much bigger solutions.