Exchange averages over coupling window instead of instantaneous values

[valentinaschueller]

Look at the documentation of the Coupled Sea Breeze case, here it says something about accumulating fluxes in time. Right now, I am computing the instantaneous flux at the surface at each coupling time step, while I should actually do a proper averaging across the whole coupling window.