8.15 Convective fluxes

field	unit	horizontal	vertical	temporal	read time
entrainment updraught	kg/cell/s	cells	full levels	[21,03], [03,09], ..	21, 03, ..
entrainment downdraught	kg/cell/s	cells	full levels	[21,03], [03,09], ..	21, 03, ..
detrainment updraught	kg/cell/s	cells	full levels	[21,03], [03,09], ..	21, 03, ..
detrainment downdraught	kg/cell/s	cells	full levels	[21,03], [03,09], ..	21, 03, ..
cloud base	level	cells	-	[21,03], [03,09], ..	21, 03, ..
cloud top	level	cells	-	[21,03], [03,09], ..	21, 03, ..
cloud level of free sinking	level	cells	-	[21,03], [03,09], ..	21, 03, ..

See section on Convective transport in the TMPP manual for how computed from raw data.

Details:

• Entrainment/detrainment updraughts/downdraughts are constant in intervals [21,03], [03,09], [09,15], and [15,21] .
• 3D arrays have only 'lmax_conv' levels; for CBM4 version with 25 levels (troposphere) lmax_conv=19 .
• Grid cells are optionally combined by area-average (updr/downdr in kg/m2/s).
• Layers are combined by summation.
• Convective cloud base, top, and level of free sinking are computed from the detrainments.
• Unit conversion entr/detr updr/downdr from kg/m2/s to kg/cell/s .
• A scale factor 'czeta' is available for adhoc testing; set to 1.0 by default.
• Netto zero tracer transport by either updraught and downdraught is ensured in each column by adding initial netto difference between total entrainment and total detrainment to the entrainment at the level where it reaches a maximum.

To be improved:

• lmax_conv should not be hardcoded but based on pressure.

Question: In meteo.F90 the netto tracer transport by updraught (downdraught) is forced to zero by adding the initial difference between entrement and detrement to the level with largest entrement.

• Couldn't this lead to negative entrements ?
• In the updraught part PB changed the initialisation of the loop over levels; shouldn't the same change be applied to the downdraught part ?