Continuation of cmorization of deglaciation runs (Kapsch et al. 2022)

Cmorization of the remaining variables of deglaciation simulations (pmu0179-pmu0212) - see earlier cmorization for metadata!

Variables to be cmorized are given in the files:

MPI-ESM1-2_echam6_PalMod2_mapping.txt MPI-ESM1-2_jsbach_PalMod2_mapping.txt MPI-ESM1-2_mpiom_PalMod2_mapping.txt

To be clarified: Ocean surface heat flux, ocean surface freshwater flux and sea-area percentage (sftof) as well as surface runoff (mrros) and total runoff (mrro).

assigned to @k204227

Update after checking the variable list submitted for CMORization/publication of additional cmip6 variables in the model output (deglaciation runs).

ECHAM output is missing the following requested variables

1. Variable hurs (code 55)
2. Variable huss (code 54)

MPIOM 3d output be copied to the data pool for following variables

1. Variable so
2. Variable thetao
3. Variable vo

assigned to @m300467 and unassigned @k204227

I confirm that the variables hurt and huss were not written out in these model runs. Hence, they can be neglected for these simulations.

3d mpiom model output for the variables so, thetao, uo and vo is unpacked and available on the pool. Note, that the mapping of the variables differs to the cmip6 file that I attached before and should be adapted for our model version.

vo (cmip6) = vke (mpi-esm-cr) uo (cmip6) = uko (mpi-esm-cr) wo (cmip6) = wko (mpi-esm-cr) thetao (cmip6) = tho (mpi-esm-cr) so (cmip6) = sao (mpi-esm-cr)

Other relevant variables are, which are used within cmip6 and have not changed names in the model:

umo Ocean Mass X Transport vmo Ocean Mass Y Transport wmo Ocean Mass Z Transport

unassigned @m300467

assigned to @m300467

unassigned @m300467

assigned to @k204227

Based on discussion with Marie on 19.05.2022, following variables to be added to mapping (PalMod-specific list, from excel sheet),

1. Ocean surface heat flux,
2. Ocean surface freshwater flux
3. Sea-area percentage (sftof)

Following variables to be added to CMOR scripts (existing CMIP6 variables? to be checked)

1. wo = wko (vertical component of ua/va)
2. umo = Ocean Mass X Transport
3. vmo = Ocean Mass Y Transport
4. wmo = Ocean Mass Z Transport
5. mrro (table Lmon)
6. mrros (table Lmon)

@m300467 For Ocean surface heat flux I can find several CMIP6 variables:

• hfx - ocean heat X transport
• hfy - ocean heat Y transport
• hfsso - ocean downward sensible heat flux
• hflso - ocean downward latent heat flux

the latter two are not available in MPIOM afaik. Is this what you had in mind? How can this be combined into one variable?

Regarding the ocean surface freshwater flux we already have the variable wfogiwosi. The difference would be that now also the contribution from sea ice is taken into account? This would then be the standard wfo variable https://clipc-services.ceda.ac.uk/dreq/u/baa63578-e5dd-11e5-8482-ac72891c3257.html ?

What shall be the frequency of sea area percentage? For now I added a time invariant and a decadal version.

The other 6 variables could be added directly from the CMIP6 tables.

hfx and hfy are the variables I had in mind. Hence, it would be great if you could include them.

For the ocean surface freshwater flux the variable wfogiwosi is sufficient. I could not find code 78 in the model output, hence, we do not seem to have an equivalent to the CMIP6 wfo output in our model version.

Sea area percentage has a decadal resolution for the transient model simulations.

I cmorized the ocean 3D variables

• Omon: so, thetao, thkcello, umo, uo, vmo vo, volcello, wmo
• Ofx: thkcello, volcello They are located at /work/bk1192/k204212/palmod_secondBatch/transient-deglaciation-prescribed-glac1d_r1i1p1f1-CR/archive/ (and similar for the other 5 runs), both on mistral and levante.

@m300467 @k204227 Please have a look. Before a publication I will need to make a new release of the applied CMOR tables and update the table_info attribute in the netCDF files.

Thanks for the efforts. The values look reasonable.

assigned to @k204212 and unassigned @k204227

assigned to @k204227 and unassigned @k204212

assigned to @k204212 and unassigned @k204227

@m300467 I have a few questions regarding the remaining variables (as linked in the txt files above):

• jsbach:
  • fx - mrsofc(field capacity) , rootd (root depth), sftgif (land ice fraction), sftlf (land fraction): Should there be a decadal version of these variables next to the fx-version that holds the state at the experiment start?
  • The files /work/mh0110/m211003/mpiesm-1.2.00p1/experiments/<simulation>/restart/topo/jsbach_T31GR30_11tiles_5layers_natural-veg_*k.nc on mistral cannot be found under the same path on levante yet. They probably have to be moved from /work/mh0110/from_mistral/m0110/....
• mpiom: The following variables do not exist in the palmod output in the way they were mapped for CMIP6. To be able to cmorize these variables, I would need the mapping information for PalMod:
  • fx - basin: for CMIP6, the basin variable can be found in /pool/data/MPIOM/input/r0010/GR15/GR15_basin.nc. That file does not exist for the mpiom grid resolution used in PalMod. Is this variable time invariant for PalMod as well?
  • fx - sftof (sea area percentage): in CMIP6: weto (code 172) in the mpiom fx stream. In PalMod this variable has decadal resolution, but I cannot find a code 172 in the PalMod output.
  • sltbasin: for CMIP6, sltbasin consists of the three variables global_sltbasin;atlantic_sltbasin;indopacific_sltbasin in the data_moc_mm stream (codes: 1112, 1113, 1114).
  • snd (snow depth): in CMIP6, there exists no mapping for this variable. It is more likely a jsbach or echam variable?!
  • hfds (surface downward heat flux): in CMIP6: hfds (code 170) in the data_2d_mm stream.
  • tauuo and tauvo (surface downward stress): in CMIP6: tauuo and tauvo (codes 548 and 549) in data_2d_mm stream.
  • simass (sea ice mass per area): in CMIP6: simass (code 1013) in the data_2d_mm stream
  • sisnconc (snow area percentage): in CMIP6: sisnconc (code 1015) in the data_2d_mm stream
  • sithick (sea ice thickness): in CMIP6 calculated with sivol/siconc (codes 13 and 15). In PalMod, code 13 corresponds to the variable sictho and not sivol. Is the recipe still fine (i.e., code 13 divided by code 15)?
  • siu (sea ice velocity, X): in CMIP6: siusic/sicu (codes 1022 and 1023) in data_2d_mm stream
  • siv (sea ice velocity, Y): in CMIP6: sivsic/sicv (codes 1024 and 1025) in data_2d_mm stream

Hi Martin @k204212 ,

jsbach:

• the variables should also be saved as decadal variables, as they are changing throughout the simulation (certainly for sftgif, sftlf).
• the files are now copied and should be available on the same path on levante.

mpiom:

• fx - basin: temporally changing but we do not written out - however we do not need it in this case, as moc, htra and wtra are already calculated.
• fx - sftof (sea area percentage): weto can be calculated with cdo -setvar,weto -setcode,172 -div sao sao
• sltbasin: not available for our simulations
• snd (snow depth): sicsno is the snow depth on sea ice, sn is the snow depth on land (can be found in the echam6_BOT files)
• sisnmass: is sicsno*0.3
• hfds (surface downward heat flux): variable flum in the data_2d_mm stream.
• tauuo and tauvo (surface downward stress): aofltxwo and aofltywe in data_2d_mm stream.
• simass (sea ice mass per area): cdo mulc,0.9 sictho
• sisnconc (snow area percentage): does not exist
• sithick (sea ice thickness): can be calculated with cdo div sictho sicomo
• siu (sea ice velocity, X): sicuo in data_2d_mm stream
• siv (sea ice velocity, Y): sicve in data_2d_mm stream

Cheers, Marie

@m300467

fx - sftof (sea area percentage): weto can be calculated with cdo -setvar,weto -setcode,172 -div sao sao

sao seems to be a 3D variable, can also a 2D variable be used for this, like sos?

Updating here the current status:

• "2nd batch" (3D ocean variables) availabe at /work/bk1192/k204212/palmod_secondBatch/transient-deglaciation-prescribed-glac1d_r1i1p1f1-CR/archive/PalMod2/ and already approved (beside areacello / volcello)
• 100a output of "3rd batch" (remaining jsbach/echam/mpiom variables) is available at /work/bk1192/k204212/palmod_thirdBatch/transient-deglaciation-prescribed-glac1d_r1i1p1f1-CR/archive/cmor_2001-2100/ and awaiting feedback, areacello and volcello yet missing
• sos can be used to calculate fx_sftof (sea area percentage)
• the ocean grid area is calculated consistent with CMIP6:
  • generate griddes
  • remove the problematic halo, then calculate gridarea
  • recreate the halo, reset the griddes to the one previously generated
  • Now we have the same grid as for all other variables, but the gridarea for a grid with properly defined halo
  • we could add a comment to each ocean variable "The data is provided on a curvilinear grid that contains a one grid cell wide halo, i.e. the first and last column of the grid are duplicates and have to be removed before conducting a calculation of areal averages or a remapping of the data."?
  • areacello has to be retracted and republished (to be consistent with CMIP6)
  • volcello has to be created again (to be consistent with CMIP6)
• there are problems to generate ps_Amon (the append mode, i.e. concatenation, does not work)
  • for the moment we use ncrcat to concatenate files after cmorization, on the longer term we hope for a fix in cdo cmor

Hi Martin, thanks for the efforts. A quick check of a couple of random output variables of the third batch of data shows that the data seem reasonable. The comment to the ocean variables seems fine and clear to me. If there are any more problems please let me know. Cheers, Marie

mentioned in issue #9