The falsifiability of weather reports is part of their strength, as you said.

A prediction that is disagreed with can be for the better or worse. Same with a prediction that is agreed with. Same with a prediction that is understood. But a prediction that is misunderstood, that generates false positives, will not be for the better. Thus a priority in prediction is not to be misunderstood. Similar to what you said.

Both insights from Karl Popper.