Cloud seeding and snow

It’s the stuff of charlatan rainmakers of old — when the weird dance doesn’t cut it, try firing some fancy pyrotechnics skywards and hope for a lucky cloudburst. Those pyrotechnics are not so very far from what Snowy Hydro has been up to for over a decade now, hoping to increase winter snowfall over our alps. But to what effect?

Of course the Hydro are no charlatans, and their cloud seeding program — which progressed from research to a routine operation in 2013 — can be shown to do better than luck. Just not by much.



Cloud seeding is not a new idea. I recall my father, a CSIRO research scientist, lauding Australia’s premier research establishment’s heroic early efforts way back in the 1960s. These days cloud seeding — or “precipitation enhancement” — seems to have made it to the mainstream, but not without a lingering whiff of scepticism and derision. The basic problem is that cloud physics is still a long way from properly understood. It’s the single largest area of uncertainty in global climate modelling.

Snowy Hydro’s cloud seeding is of the best understood kind — glaciogenic seeding of orographic clouds. Essentially, you introduce ice-initiating microparticles into a supercooled water vapour cloud, hoping to generate ice crystals (eventually snowflakes) via the Bergeron-Findeison mechanism. That exploits the different vapour pressures of ice and water at the same sub-zero temperature, allowing an introduced ice nucleus to grow rapidly at the expense of nearby supercooled water droplets. But … how can those two co-exist? That is simply because phase changes do not have to coincide with the boundaries of the stable ranges of material phases. Water doesn’t necessarily change instantly to ice at 0°C. Instead, in the absence of a suitable trigger for change, it can remain liquid in a metastable or “supercooled” state to a much lower temperature – as low as -35°C for cloud droplets. An orographic cloud forming in a westerly airstream lifting over our alps is a likely place for such conditions.


Cloud seeding burner (Snowy Hydro photo)

Snowy Hydro’s approach

Snowy Hydro’s cloud seeding is delightfully simple and, presumably, cost-effective. The seeding chemical is a very old one — silver iodide, a nearly insoluble crystalline salt with a similar hexagonal crystal structure to that of water ice, making it an excellent ice crystal seed. It’s also relatively cheap (despite the silver — it was the active ingredient in all that photographic film) and non-toxic. Importantly it is soluble in a flammable organic solvent (acetone — in the presence of sodium nitrate), which can readily be burnt to produce a smoke of trillions of ultra-fine (~60 nm) silver iodide particles.

Unlike Australia’s other operational cloud seeding program over western Tasmania (conducted by Hydro Tasmania), Snowy Hydro does not use aircraft to disperse the agent. Instead they use 23 ground-based burners strategically located along the western flank of the NSW alps. In the right conditions, the units are remotely triggered to burn silver iodide solution in LPG-fired burners, which generate sufficient heat to lift the resulting silver iodide smoke high into the airstream, where it rapidly mixes into the developing orographic cloud.

We know that part, at least, works well, because the silver plume (or added tracers) can be measured at miniscule concentrations in the resulting snow fallout, and reconciled with the addition. The big question is, did the silver make the snow, or would it have happened anyway?



The first and most obvious thing to say is that Snowy Hydro’s cloud seeding clearly does not have a large effect, because if it did, they’ve been at it long enough (since 2004) and hard enough for that to be obvious by now in the snow depth records. But it’s not. For example here’s the moving average Spencers Creek¹ season peak snow depth trace, which if anything has been a little below trend of late:

Spencers Creek peak snow depth moving average and trend

Spencers Creek peak snow depth moving average and trend

But that definitely does not mean that it has no effect. Unlike many such enterprises, Snowy Hydro designed and implemented a thorough monitoring scheme to assess effectiveness, and reported the results in peer-reviewed journal articles.2,3  The headline finding was a “14% increase in precipitation from selected suitable storms”. All good, except note that’s not 14% more snow depth, not even 14% more total snowfall, it’s 14% more measured precipitation from suitable seeded events. And it turns out that you only get to that answer by excluding the experiments you didn’t like (“overall target (not) effectively covered”) … otherwise it’s only a 7% gain from the seeded events, with, well, 24% probability that the observed gain was due to chance alone.

My view? Snowy Hydro’s cloud seeding is very likely having some modest effect, but don’t expect it to be enough to counteract the global warming downtrend. That is much more vigorous and robust.



  1. Spencers Creek near Charlotte Pass, midway between Perisher Valley and Thredbo, New South Wales, Australia; data from Snowy Hydro Limited.
  2. Manton, M. J., Warren, L., Kenyon, S. L., Peace, A. D., Bilish, S. P., & Kemsley, K. (2011). A confirmatory snowfall enhancement project in the Snowy Mountains of Australia. Part I: Project design and response variables. Journal of Applied Meteorology and Climatology, 50(7), 1432-1447.
  3. Manton, M. J., & Warren, L. (2011). A confirmatory snowfall enhancement project in the Snowy Mountains of Australia. Part II: Primary and associated analyses. Journal of Applied Meteorology and Climatology, 50(7), 1448-1458.