Our attempts to 'seed' clouds with extra snow or rain may actually be working, a new study suggests, although perhaps not quite as much as some had hoped.

It may sound like something straight out of science fiction, but cloud seeding has been around since the 1940s. Despite the fact that this form of 'weather modification' is routinely used throughout the United States and elsewhere, its reputation is about as dubious as the precipitation it supposedly triggers.

Today, scientists are still not sure if cloud seeding actually works outside the lab. Previous studies in the real world have produced mixed results, and the ones that do show benefits are often not reproducible. Besides, in many of these studies and real-world situations, it's difficult to determine what's natural rain and what's not.

Now, the first attempt to accurately measure cloud seeding, funded in part by the US National Science Foundation (NSF), has found a flight's worth of seeding produces barely enough snow to dust your eyelashes.

That said, if the cloud hadn't been seeded, it wouldn't have produced any precipitation at all.

"Everyone you talk to will say even if you can generate a little bit more snow, that helps us in the long run," says atmospheric scientist Katja Friedrich from the University of Colorado Boulder.

Over half a century ago, scientists discovered that tiny 'seeds' of dry ice could provide a scaffold for water molecules, allowing them to line up and crystallise (or freeze) at lower temperatures than normal. Silver iodide was later determined to be a more efficient material.

Since then, the idea that this knowledge could be harnessed to increase water supply has led many nations like Australia to invest in the process, despite a lack of solid evidence that it actually works, among other unresolved concerns.

Unlike previous studies, which are often unable to watch the seeded cloud as it evolves, Friedrich and her colleagues were able to use a radar dish to accurately calculate the volume of snow produced through cloud seeding.

"We have airplanes now with radars on them that can see things that we could never see before," atmospheric scientist Bob Rauber from the University of Illinois told local news outlet The Center Square.

"We have radars that operate at much higher power and frequency that we can see things we couldn't see before."

During winter, they had a plane inject particles of silver iodide into a natural cloud passing overhead in western Idaho. They then used the radar and other precipitation measurements to 'watch' the cloud's water vapour thicken and grow heavy with droplets.

"If everything goes according to plan, the water droplets will begin to freeze around the aerosols, forming snow," explains Friedrich.

And things don't always go to plan. Weather is a fickle thing, and nearly impossible to reproduce given the sheer number of factors that would have to be taken into account.

Later that month, for instance, wind conditions sent precipitation from the team's seeded cloud farther downwind than intended. Another time, the snow fell too quickly.

On that very first cold January day, however, the team says snow fell for just over an hour, dusting 2,330 square kilometres (900 square miles) in about a tenth of a millimetre of snow.

Including that day's artificial dump and the two other occasions, the authors say they seeded enough water to fill 282 Olympic-sized swimming pools.

As the clouds passed over their ground-based instrument, they say 0.2 mm fell in five minutes. Taking average snow falls into account before and after the fall, they figured half of that could be attributed directly to the seeding. Compared to the region's natural variability, that means "the increase due to seeding is three to four times greater".

"We tracked the seeding plume from the time we put it into the cloud until it generated snow that actually fell onto the ground," says Friedrich.

"We can now finally put a number on how much water we can produce through cloud seeding."

So far, estimates on the efficiency of cloud seeding have varied considerably, anywhere from zero to 50 percent additional snowfall.

The new findings suggest the answer lies somewhere in the middle, but for some critics of the practice, that may not be enough.

Friedrich herself says she doesn't know how useful cloud seeding will be when it comes to increasing water supply, but she says her findings get us closer to being able to make that cost-benefit analysis.

Despite recent interest and growth, the use of cloud seeding remains very controversial. Some say it isn't worth the resources, if it only increases rainfall to around 20 percent, and if it's done incorrectly, there might be risks.

In some cases, it might reduce rainfall downwind, where the cloud would have normally let go of its rain, if not for being prodded by us humans. Whereas others claim it increases precipitation downwind.

Other concerns centre around the particles themselves, which despite being natural and nearly impossible to detect in the snowpack, may theoretically pose a risk if cloud seeding intensifies beyond reasonable levels.

We clearly need more research to figure out how we can use this practice more precisely, and also predict what its environmental, social, and legal consequences will be.

The study was published in PNAS.