On a chilly clear day in January, a team of scientists turned their heads to the sky and watched as snowflakes began to fall above them in western Idaho. There was no snow predicted on the weather reports – well, at least, by meteorologist forecasts. 

This snow was created by scientists firing a series of flares from an aircraft to inject silver iodide (AgI) particles into a natural cloud floating by. The experiment, called SNOWIE (Seeded and Natural Orographic Wintertime Clouds: The Idaho Experiment), happened two more times over the course of the month. 

The team led by the University of Colorado Boulder triggered the snow via a technology called "cloud seeding", a much-discussed, much-debated method of increasing water supply to arid regions. To clarify the many uncertainties of such an approach, the team took to the difficult task of measuring artificially provoked precipitation. 

The study, sponsored by the National Science Foundation, is the one of the first to "show the entire chain of events during three cases from the time we put the seeding material into the cloud, as it converts supercooled liquid into snow, and how the snow falls onto the ground," said Katja Friedrich, an associate professor at the University of Colorado Boulder, to IFLScience.

In all, the seeding events produced enough water to fill around 282 Olympic-sized swimming pools. That may sound like a lot, and in some ways, it is, however the 2,330 square kilometers (900 square miles) the dusting of snow landed upon came out to be about a tenth of a millimeter deep. The findings are published in the journal Proceedings of the National Academy of Sciences.

So how does cloud seeding work? Silver iodide particles are used as "seeds" that are injected into the clouds – the little bits of material provide a surface for water molecules to latch onto and form ice crystals around, eventually producing rain or snow. However, proof that such a method works (and is worth the effort) is controversial. Studies have yielded conflicting or inconclusive results, with specific measurements difficult to attain: Is the additional precipitation natural or man-made? The uncertainties of weather make cloud seeding a less-than-ideal uncontrolled experiment. 

"Previous studies used statistical comparisons between seeded and non-seeded events and times as well as target and control areas – the big caveat here is are the conditions the same? Some studies reported enhancements in snowfall due to cloud seeding that range between 0.5 and 2 mm h-1 but these are only point measurements and we cannot say how much water is being generated over the entire catchment," said Friedrich.

However, "This paper quantifies the amount of snow produced over an entire catchment."

Another question that rises to the surface is whether there are environmental costs of such an inference to natural weather patterns. A previous report suggested that preliminary hazards seem low. Yet this isn’t to say there aren't any, just that the evidence for a definitive answer is not there yet. It's also important to note that the technology is powerless without clouds and their location matters too. Concerns that silver iodide can make people sick have so far shown negligible effects at the levels used to seed clouds.

Pilots have previously modified weather patterns to stop hail destroying crops in Alberta, Canada. Such measures are becoming more popular, but the research is still limited in what it can say are the final effects. For states such as Idaho and Colorado, rivers that are used as water for districts depend on the amount of snowfall they receive. 

"The clouds are seeded with silver iodide, which will end up in streams and rivers. In general, the cloud seeding entities have to comply to environmental regulations, like every other company," noted Friedrich. "As far as I know the amount of silver ending up in the ground water is low and the same order of magnitude as the natural silver background concentration. However, this is not my area of expertise."

In addition, "the study only looked at three cases. Water managers want to know how much water is being produced over an entire water year or winter season. That would be the next step that we want to address with numerical modeling."