If your goal is people growing more food relative to baseline, you should subsidize food (which we do, intensively).
If your goal is people living in flood zones, you should subsidize living in flood zones (e.g. by insuring it, which we also do).
We need an answer to the latter question in this case, not the former. And one is readily at hand: the political interest of those who live in flood zones. Maybe there's some other justification, but I'll confess I'm skeptical that this outcome is globally utility maximizing relative to alternative policies.
I'm sure other people have thought this through much better than what I just said. But it seems to me to be a reasonable argument.