Move leaf_c calculation after cohort area and number adjustment in SP mode calcuations

[glemieux]

In the course of dealing with some testing for #1024, I noticed that we are currently updating the cohort area and number after we calculate the leaf_c in calculate_SP_properties:

fates/biogeochem/EDPhysiologyMod.F90

Lines 1929 to 1953 in 4f41112

	leaf_c = leafc_from_treelai(tlai, pft, c_area, cohort_n, canopy_layer, vcmax25top)
	! check that the inverse calculation of leafc from treelai is the same as the
	! standard calculation of treelai from leafc. Maybe can delete eventually?
	check_treelai = tree_lai(leaf_c, pft, c_area, cohort_n, canopy_layer, &
	canopylai, vcmax25top)
	if (abs(tlai - check_treelai) .gt. 1.0e-12) then !this is not as precise as nearzero
	write(fates_log(),*) 'error in validate treelai', tlai, check_treelai, tlai - check_treelai
	write(fates_log(),*) 'tree_lai inputs: ', pft, c_area, cohort_n, &
	canopy_layer, vcmax25top
	call endrun(msg=errMsg(sourcefile, __LINE__))
	end if
	! the carea_allom routine sometimes generates precision-tolerance level errors in the canopy area
	! these mean that the canopy area does not exactly add up to the patch area, which causes chaos in
	! the radiation routines. Correct both the area and the 'n' to remove error, and don't use
	! carea_allom in SP mode after this point.
	if (abs(c_area - parea) .gt. nearzero) then ! there is an error
	if (abs(c_area - parea) .lt. 10.e-9) then ! correct this if it's a very small error
	oldcarea = c_area
	! generate new cohort area
	c_area = c_area - (c_area - parea)
	cohort_n = cohort_n*(c_area/oldcarea)

I believe we should defer the leaf_c calculation until after the area and number adjustments since they are inputs to leafc_from_treelai.

See discussion starting from #1024 (comment) for more context.