GCHP mass flux wrong moisture correction

[yuanjianz]

Your name

Yuanjian Zhang

Your affiliation

WashU

What happened? What did you expect to happen?

GCHP/src/GCHP_GridComp/GCHPctmEnv_GridComp/GCHPctmEnv_GridCompMod.F90

Lines 873 to 885 in e499969

	if ( correct_mass_flux_for_humidity > 0 ) then
	MFX_EXPORT = dble(temp3d_r4) / ( 1.d0 - SPHU0_EXPORT )
	else
	MFX_EXPORT = dble(temp3d_r4)
	endif
	call MAPL_GetPointer(IMPORT, temp3d_r4, 'MFYC', RC=STATUS)
	_VERIFY(STATUS)
	if ( correct_mass_flux_for_humidity > 0 ) then
	MFY_EXPORT = dble(temp3d_r4) / ( 1.d0 - SPHU0_EXPORT )
	else
	MFY_EXPORT = dble(temp3d_r4)
	endif

Humidity is positive so 1-SPHU is smaller than 1, so dryMF = MF/(1-SPHU) is bigger than MF, which does not make sense to me. Shouldn't mass flux from GMAO archives be multiplied instead of being divided by 1-SPHU?

[lizziel]

This makes sense to me. 1/(1-sphu) is the same as kg total / kg dry, but it would make sense to convert by instead multiplying by kg dry/kg total. When you change it do you see improvements in comparisons with wind runs?

[yuanjianz]

Theoretically, this bug fix is reasonable and a trivial change. But the test reminds me of the previous comparison between mass flux and wind. With / without this fix, the advection in mass flux is much less vigorous than in wind. Together with my recent massflux v.s. wind fullchem benchmark showing unreasonble surface aerosol concentration, I think it is necessary to continue the investigation in #342 with help from @sdeastham .

[lizziel]

Using mass fluxes in GCHP is still experimental, even with this update. Do you want this to be merged or wait until more testing is done?

[yuanjianz]

I think this can be merged first since it is not intercorrelated with other mass flux issues.