The idea is that a flooded plant needs to pump water out of space
surrounding the roots as quickly as possible. Epinasty then does this
through ramped up tranpiration induced by a higher water pressure from
a lever pump action. (Maybe downward directed leaves transpire more
quickly in the wind even without pump action because when they become
perpedicular to it -to the wind- or when the plant makes sure to
epinastically grow perpendicular to the most common prevailing wind -
this makes transpiration faster than if the leaves are parallel to the
wind as the are on the most part during normal conditions.)
I'm sure this has been thought of and discussed before even by
Aristotle and others of his ilk.