How we determine if a storm is seedable

Water droplets in the atmosphere can remain liquid well below freezing when there is nothing present for ice to form on. That state — supercooled liquid water — is what glaciogenic seeding targets. Introducing a material with a crystal structure similar to ice gives those droplets a surface to freeze onto. The resulting ice crystals grow at the expense of the surrounding droplets and can fall out as snow that the cloud would otherwise have carried away.

If a cloud holds no supercooled liquid water, seeding has nothing to act on. This is the single most common reason a storm is not a candidate, and it is not always obvious from the ground: a storm can be producing heavy snow and still be glaciated — already full of ice — with little liquid left to convert.

Cloud temperature bounds the opportunity on both sides. Too warm and there is no supercooling to work with. Too cold and the cloud has typically already glaciated naturally, so additional nuclei add little. Operational programs therefore work within a cloud-temperature band rather than treating all cold clouds as equivalent, and they consider where in the cloud depth that band sits — because the material has to reach it.

The practical question is not "is it cold?" but "is there a layer at a usable temperature that also contains liquid water, and can we place material into that layer?"

A seedable layer that exists for ten minutes is not an operational target. Sustained upward motion — usually orographic lift as a storm is forced over a mountain range, sometimes frontal lift — is what continuously regenerates supercooled water and gives a mission a window long enough to matter.

Moisture supply matters equally. Seeding redistributes and accelerates precipitation from water that is already in the air. A dry storm with a good temperature profile is still a poor candidate.

Material released into a cloud travels with the wind while ice crystals grow and fall. The time between nucleation and snow reaching the ground means the release point and the intended catchment are not the same place. Wind speed and direction through the seeding layer determine where the effect, if any, lands.

A storm can be physically seedable and still be rejected because the trajectory carries the effect outside the target watershed. For a water district, precipitation outside the contributing basin is not the outcome being paid for. This is a large part of why targeting precision matters — see the comparison of delivery methods.

• Airspace. Coordination and authorizations have to be in place for the specific volume and time window.
• Icing and turbulence. The conditions that make a cloud seedable are the same conditions that ice airframes. Platform limits are a hard constraint, not a preference.
• Surface conditions. Launch and recovery sites in mountain winter weather have their own envelope.
• Downstream conditions. Avalanche risk, flood conditions, and program suspension criteria can make an otherwise good storm a no-go.

In practice, the decision is layered. Forecast model output identifies candidate storms days out. Soundings, satellite-derived cloud properties, and radar refine the picture as the storm approaches. In the final hours, real-time observation determines whether the seedable layer exists where the model said it would, and the mission either proceeds or stands down.

Standing down is a normal outcome. A program that flies every storm is not exercising judgment; it is spending resources on clouds that cannot respond. More on the instrumentation behind these calls in how atmospheric data guides cloud seeding, and on the conditions common in the Intermountain West in cloud seeding in Utah.