“Predict flooding” is a uselessly vague term for you to use.
Indeed it is all about the granularity. That's the difference between making money from the usefulness of your predictions rather than just making money convincing others your predictions are useful.
Like saying that when my mother-in-law got her terminal cancer diagnosis, the oncologist didn’t predict her death because you have in mind some level of precision for a prediction that she would die that meant that the terminal diagnosis wasn’t a good prediction of her death.
I'm sorry about your mother-in-law, but regardless, everybody dies eventually, while cancer definitely narrows the scope of that prediction. Doctors will often give patients some range of assumed life expectancy based on their condition. If these folks can say flooding is likely in 3 years or 10 years that is useful. OTOH if they say it may be more likely to flood one day that is as useful as knowing everybody dies. I would not pay for that.