Scaling of cohort crown area in the output files?

[mpaiao]

I am adding a new crown area variable to the output (by size class and PFT). I was trying to build on existing variables in FatesHistoryInterfaceMod.F90, but I am somewhat confused. When defining crown area by size class, the cohort variable is scaled by 1 / m2_per_ha:

fates/main/FatesHistoryInterfaceMod.F90

Lines 3386 to 3387 in 5551a70

	hio_crown_area_canopy_si_scls(io_si,scls) = hio_crown_area_canopy_si_scls(io_si,scls) + &
	ccohort%c_area / m2_per_ha

fates/main/FatesHistoryInterfaceMod.F90

Lines 3535 to 3536 in 5551a70

	hio_crown_area_understory_si_scls(io_si,scls) = hio_crown_area_understory_si_scls(io_si,scls) + &
	ccohort%c_area / m2_per_ha

However, when defining crown area by PFT, the cohort variable is scaled by AREA_INV:

fates/main/FatesHistoryInterfaceMod.F90

Lines 3041 to 3048 in 5551a70

	hio_crownarea_si_pft(io_si, ft) = hio_crownarea_si_pft(io_si, ft) + &
	ccohort%c_area * AREA_INV
	if (ccohort%canopy_layer .eq. 1) then
	! Update PFT canopy crown area
	hio_canopycrownarea_si_pft(io_si, ft) = hio_canopycrownarea_si_pft(io_si, ft) + &
	ccohort%c_area * AREA_INV
	end if

Is there any reason for the different scaling, or should both be scaled by AREA_INV?

[ckoven]

@mpaiao I think it is the same conversion? i.e. multiplying by AREA_INV is the same as dividing by m2_per_ha. I think the implied meaning of the unit conversion is very slightly different though; AREA and AREA_INV are what set the nominal size of a FATES site to be 1 hectare, but in theory at least that could vary, and if it did, then the m2_per_ha would no longer be the correct unit conversion. So in principle AREA_INV would seem to be a very slightly safer unit conversion. Also seems weird that the two constants live in different modules.

fates/main/EDTypesMod.F90

Lines 76 to 77 in 5551a70

	real(r8), parameter, public :: area = 10000.0_r8 ! Notional area of simulated forest m2
	real(r8), parameter, public :: area_inv = 1.0e-4_r8 ! Inverse of the notion area (faster math)

fates/main/FatesConstantsMod.F90

Lines 223 to 224 in 5551a70

	! Conversion factor: m2 per ha
	real(fates_r8), parameter, public :: m2_per_ha = 1.0e4_fates_r8

[mpaiao]

@ckoven Thanks for clarifying. In the short term, I will replace m2_per_ha with AREA_INV when I submit the PR for the additional crown area variable, if anything, to be consistent with the other crown area conversions.

But maybe we can have a long-term discussion on whether or not to even have a nominal area at some point. If nplant is defined as a true demographic density (stems m⁻²), then the site area is not really needed for anything, and it would reduce the number of multiplication operations in the code.

[ckoven]

@mpaiao agreed, that would be nice (but also will probably touch everything!)