Radiative fluxes at top-of-the-atmosphere and top-of-the-model in NorESM2 output

[DirkOlivie]

Introduction
Currently, in NorESM2 some radiative fluxes in the output are top-of-the-atmosphere (TOA) fluxes, whereas others are top-of-the-model (TOM) fluxes. Because NorESM2 has a low model top at around 3 hPa, there is a considerable difference in diagnosing a flux at TOA compared to TOM : global annual mean fluxes can differ up to 1.5 W/m2.

For CIMP6, fluxes have to be diagnosed at TOA. Therefore some fluxes had to be converted, e.g. : rlut= FSNTOA - FSNT + FLNT.

Some radiative fluxes have names and descriptions in the output files which are confusing. E.g., FLNTCDRF, FLNT_DRF, FSNTCDRF, or FSNT_DRF.

Suggestion

• Output in the future versions of NorESM radiative fluxes both at TOA and TOM
• Give new name and description to some fluxes, to make clear what they actually are. E.g., :
  • FLNTCDRF becomes FLNTCSAF
  • FLNT_DRF becomes FLNTAF
  • FSNTCDRF becomes FSNTCSAF
  • FSNT_DRF becomes FSNTAF
  • Similar renamings for surface fluxes FSDSCDRF, FSDS_DRF.FSNSCDRF, FSNS_DRF, and FSUS_DRF.
  • Renaming for FSUTADRF.

Complication

• Diagnostics package : some of these fluxes are (possibly) used in the atmospheric diagnostic package.
• Cmorization : some of these fluxes are certainly used during the cmorization.

If these modifications are useful, one should think about from which version onwards of NorESM and CAM-Nor, these modifications can be implemented without causing too many problems.

[adagj]

I agree that we indeed should change the variable names as well, but here is a suggestion for a description of the AEROFFL and AEROCOM variables:

#ifdef AEROFFL
call addfld ('FSNT_DRF',horiz_only, 'A','W/m^2','Clean sky (aer_tau=0) net solar flux at top of model')
call addfld ('FSNTCDRF',horiz_only, 'A','W/m^2',' Clean (aer_tau=0) and Clear sky net solar flux at top of model' )
call addfld ('FSNS_DRF',horiz_only, 'A','W/m^2 ','Clean sky (aer_tau=0) net solar flux at surface' )
call addfld ('FSNSCDRF',horiz_only, 'A','W/m^2 ','Clean (aer_tau=0) and Clear sky net solar flux at surface ' )
call addfld ('QRS_DRF ',(/'lev'/), 'A','K/s ','Clean sky (aer_tau=0) Solar heating rate')
call addfld ('QRSC_DRF',(/'lev'/), 'A','K/s ','Clean (aer_tau=0) and Clear sky solar heating rate' )
call addfld ('FLNT_DRF',horiz_only, 'A','W/m^2 ','Clean sky (aer_tau=0) net longwave flux at top of model' )
call addfld ('FLNTCDRF',horiz_only, 'A','W/m^2 ','Clean (aer_tau=0) and Clear sky net longwave flux at top of model' )
#ifdef AEROCOM
call addfld ('FSUTADRF',horiz_only, 'A','W/m^2 ','Clean sky (aer_tau=0) upwelling solar flux at TOA')
call addfld ('FSDS_DRF',horiz_only, 'A','W/m^2 ','Clean sky (aer_tau=0) downwelling solar flux at surface')
call addfld ('FSUS_DRF',horiz_only, 'A','W/m^2 ','Clean sky (aer_tau=0) upwelling solar flux at surface')
call addfld ('FSDSCDRF',horiz_only, 'A','W/m^2 ','Clean (aer_tau=0) and Clear sky downwelling solar flux at surface')
call addfld ('FLUS ',horiz_only, 'A','W/m^2 ','LW surface upwelling flux')
call addfld ('FLNT_ORG',horiz_only, 'A','W/m^2 ','Total column longwave flux (CAM5)' )
#endif ! aerocom

[YanchunHe]

Just come up with an idea to handle new variables to create two entries for the same CMIP6 variables with old and new names, in the variable-definition namelist files, such as:

rlut= FSNTOA - FSNT + FLNT
rlut= FSNTOA_new - FSNT_new + FLNT_new

So that the old name will firstly be searched and used, if exist; then search for new names and will cmorized with new variables names; if both old and new name exist (which should be not usually happen), both the old and new variables will be cmorized but the new new will overwrite the cmorized one with the old names.

The cmorization program can naturally handle such implement with its current design. And something similar has been done for some variables.

In this way, this is fully compatible with old and new variable names without any 'if clause' in the code.

However, a bit more complicated for the diagnostics, but should be manageable.

So renaming can be handled properly during cmorization and diagnostic.

[adagj]

Also discussed in NorESMhub/noresm2cmor#312