Closing the surface energy budget from the atmospheric side using the CMIP output

[YanchunHe]

Quote A. Nummelin

Timothee noted that there was an apparent mismatch between the oceanic and atmospheric surface heat budgets beyond what one would expect from the different grids. Basically, he was looking if hfds = - hfss - hfls + rlds - rlus + rsds - rsus holds (left from ocean output, right from atmospheric output). For most models this seemed to work, but for NorESM there was a larger than expected mismatch.

I had a look at the original output, and the terms that I would expect the budget to close with (now just using the net radiative terms) are

hflx = FSNS-FLNS-SHFL-LHFLX (again, left-hand side is from the ocean output, right-hand side is from the atmospheric output)

Here the atmospheric output is scaled by the land-sea fraction. However, in the output, there is also an EFIX variable which I'm not familiar with before, and it seems to be a bit hard to judge if it should be included.

I'm not sure if the lack of EFIX in the CMORized output https://docs.google.com/spreadsheets/d/1-XOmrpOewOJ4msZ2cuQ41NRiMKgMZ5BfiNeAAFUsp3M/edit#gid=251590531 is causing the difference Timothee is seeing but would be great to hear if it should be included or not.

Could perhaps Thomas or Mats comment on this? Thanks for the help!

Zonal mean atmospheric vs ocean net surface heat flux in one month, note that differences south of 50S are due to sea ice.

[YanchunHe]

Quote J. Tjiputra

And here are the comparison for the cumulated global fluxes for NorESM1-M and NorESM2_LM for ocean grids only.
hfds vs atm-fields
where 'atm-fields' = - hfss - hfls + rlds - rlus + rsds - rsus

[YanchunHe]

I see for the cmorization, the upward longwave radiation is derived as:
rlus=FLDS+FLNS (downward plus net longwave radiation)

the upward shortwave radiation:
rsus=FSDS-FSNS (downward minus net shortwave radiation).

Probably this is wrong?

I don’t remember when was this implented as such, but probably inherited from CMIP5.

What is the direction for the net short/longwave radiation? positive upward or downward? The direction is not explicitly specified in the raw model output.

[YanchunHe]

I see for the cmorization, the upward longwave radiation is derived as:
rlus=FLDS+FLNS (downward plus net longwave radiation)

the upward shortwave radiation:
rsus=FSDS-FSNS (downward minus net shortwave radiation).

Probably this is wrong?

It looks right for the derived upward longwave and shortwave radiations

rlus=FLDS+FLNS
rsus=FSDS-FSNS

There are about -0.64 w/m2 imbalance between heat flux between air and sea between lat 50-0 degrees (0.50 w/m2 if EFIX is included).

I can't explain what should the heat budget imbalance expected.

[YanchunHe]

close it now. reopen if there are updates in the future.

[adagj]

@mvertens @adagj
We are trying to understand the discrepancy, did any of you solve/understand this?