Refactor yearly net uptake

[glemieux]

This issue came up in the context of review #623 (comment). The trim_canopy routine uses this variable to check to see if there has been activity in a leaf layer for a given year:

fates/biogeochem/EDPhysiologyMod.F90

Line 456 in 50e3c74

if (currentCohort%year_net_uptake(z) /= 999._r8)then !there was activity this year in this leaf layer.

As noted by @rgknox this type of equivalence check against real numbers should be avoided. This type of check is also present in DeallocateCohort:

fates/biogeochem/EDCohortDynamicsMod.F90

Lines 1349 to 1357 in 50e3c74

	do i=1, nlevleaf
	if (currentCohort%year_net_uptake(i) == 999._r8 .or. nextc%year_net_uptake(i) == 999._r8) then
	currentCohort%year_net_uptake(i) = &
	min(nextc%year_net_uptake(i),currentCohort%year_net_uptake(i))
	else
	currentCohort%year_net_uptake(i) = (currentCohort%n*currentCohort%year_net_uptake(i) + &
	nextc%n*nextc%year_net_uptake(i))/newn
	endif
	enddo

Given that this is slightly beyond the scope of #623, I'm breaking this out into a separate issue to be addressed later.