FAQ: Data ocean slab mode (DOCN-SOM)

[m_dvorak]

It works now, thank you! Running ./preview_namelists with the prior file path also just stalled.

 

[dbailey]

That is very strange. I guess the path is somehow being overridden ... I'll see if this is still a problem in CESM2+ ...

 

[tresamt]

Hi Dave,

I'd like to run a F1850C5 compset with SST/ICEFRAC boundary conditions derived from the climatology of a E1850C5CN compset run. I am running CESM 1.1.2. My questions are:

1) Is there an example boundary condition file available in the online repository I can look at to help me create my own?
2) Do I need to do the 'diddling' procedure using 'icesst tools' mentioned in other threads (e.g. here)? Or is this just for using observations as boundary conditions?
3) How do I point the model to the boundary condition file I created?
4) What type of run should I do, start-up, branch or hybrid ? The simulation will go from slab-ocean to fixed sst so not sure what is recommended here.
5) Is there a way to make an F1850C5CN type compset by running F1850C5 and adding a flag for 'CN' ?
6) Is it possible to do a similar F compset using boundary conditions derived from the climatology of a QSC5 compset run? (Here i'd be using another version of the model cesm 2.1.0)

Thanks for your help,
Etienne

Im also curious about this. Where can I get details about the same?

 

[tresamt]

Im also curious about this. Where can I get details about the same?

I have some specific questions too:

The SST input file for F-compset contains only SST and ice_cov as variables. The resolution being that of CAM too.

Since, CAM has monthly outputs of SST and ice_cov, can the monthly average of these variables be used as boundary conditions for F-compset (after linearly interpolating for the mid-of-month, to match with the time dimension)?

 

[dbailey]

I have some specific questions too:

The SST input file for F-compset contains only SST and ice_cov as variables. The resolution being that of CAM too.

Since, CAM has monthly outputs of SST and ice_cov, can the monthly average of these variables be used as boundary conditions for F-compset (after linearly interpolating for the mid-of-month, to match with the time dimension)?

I'm confused here. Are you saying you want to use SST and ice_cov from an existing compset run? This particular forum is about the DOCN-SOM mode or E compsets.

 

[tresamt]

I'm confused here. Are you saying you want to use SST and ice_cov from an existing compset run? This particular forum is about the DOCN-SOM mode or E compsets.

Yes, I have a E-compset run and I want to use the SST and ice_cov from that run to make a F-compset run.

 

[dbailey]

Hi Dave,

I'd like to run a F1850C5 compset with SST/ICEFRAC boundary conditions derived from the climatology of a E1850C5CN compset run. I am running CESM 1.1.2. My questions are:

1) Is there an example boundary condition file available in the online repository I can look at to help me create my own?
2) Do I need to do the 'diddling' procedure using 'icesst tools' mentioned in other threads (e.g. here)? Or is this just for using observations as boundary conditions?
3) How do I point the model to the boundary condition file I created?
4) What type of run should I do, start-up, branch or hybrid ? The simulation will go from slab-ocean to fixed sst so not sure what is recommended here.
5) Is there a way to make an F1850C5CN type compset by running F1850C5 and adding a flag for 'CN' ?
6) Is it possible to do a similar F compset using boundary conditions derived from the climatology of a QSC5 compset run? (Here i'd be using another version of the model cesm 2.1.0)

Thanks for your help,
Etienne

Did you ever get a reply to this. @hannay can you help with this one?

 

[dbailey]

Yes, I have a E-compset run and I want to use the SST and ice_cov from that run to make a F-compset run.

The trick here is that SST from CAM is not right under the sea ice. This will be the surface temperature on top of the sea ice. You can trick it by setting any temperature below -1.8C to be -1.8C. Then it should be fine.

 

[tresamt]

The trick here is that SST from CAM is not right under the sea ice. This will be the surface temperature on top of the sea ice. You can trick it by setting any temperature below -1.8C to be -1.8C. Then it should be fine.

What if I use the SST data of ice component?

 

[dbailey]

This is possible, but a bit more complex. SST is not a default variable that is output from the sea ice and it is on a different grid (gx1 Greenland dipole grid) from the atmosphere. So, it needs to be remapped. One can use the ESMF remapping tools for this purpose.

 

[tresamt]

This is possible, but a bit more complex. SST is not a default variable that is output from the sea ice and it is on a different grid (gx1 Greenland dipole grid) from the atmosphere. So, it needs to be remapped. One can use the ESMF remapping tools for this purpose.

But for the Fcompset with f19_g16 resolution, wont the input files of SST and AICE be in gx1v6 itself? So can't the output of ice component SST and AICE of E-compset f19_g16 be used for creating the input file for F, f19_g16?

 

[dbailey]

You could do that. Typically we run the sea ice and data ocean at the same resolution as the atmosphere f19_f19 say in the F compset.

 

[tresamt]

You could do that. Typically we run the sea ice and data ocean at the same resolution as the atmosphere f19_f19 say in the F compset.

I have E-compset runs at f19_g16, so was planning to run the F-compset also at the same resolution. So do you think the outputs of ICE component of E-compset corrected for time (middle of month, rather than end) be used for running the F-comspet?

Is there any special reason why f19_f19 is used for F-compset?

Thankyou for the prompt replies

 

[dbailey]

This goes back to the original standalone CAM where everything had to be on the same grid. f19_g16 should work ... hopefully the mapping files and everything will be setup correctly.

 

[tresamt]

This goes back to the original standalone CAM where everything had to be on the same grid. f19_g16 should work ... hopefully the mapping files and everything will be setup correctly.

Hi Dave,
I created the input files for F-compset f19_g16 from E-compset. Wanted to know whether I should start the F-compset case as a start-up case or hybrid/branch? I feel it should be start-up with SST forcing provided by the new files created. But wanted to get your opinion.
Thank you.

 

[dbailey]

It should definitely be a hybrid.

 

[tresamt]

It should definitely be a hybrid.

But why should it be hybrid?
If I start F1850 as a startup case with the SST forcing file created from E1850, shouldn't it work?

 

[tresamt]

It should definitely be a hybrid.

Hi Dave,
As suggested, I used SST and AICE from the ice-component of an E-compset run (last 50 years average values taken from a 100year run). I regridded from f19_g16 to f19_f19 and also applied the corrections suggested in the ncl program: ncl scripts to create SST/ICE data files to drive an F compset using B compset history files.

I made a copy of the input file of F1850- f19_f19: SSTICE_DATA_FILENAME and replaced the SST and AICE variables with the regridded values obtained from above step. Then started a hybrid run of F1850- f19_f19 from a E1850-f19_g16 run. The CO2 concentration is 450ppm in both E and F case.

The run is successful, but the radiative imbalance at TOA for the F case is high (around 2W/m2) while the same for E case was around zero.

I'm diagnosing the reasons for the same from output files. Wanted to confirm something. Since I created a single year data (averaging over 50 years), I did not make any changes to the following :

<!--"The model year that corresponds to SSTICE_YEAR_START on the data file. (integer) " -->
<entry id="SSTICE_YEAR_ALIGN" value="1" />

<!--"The first year of data to use from SSTICE_DATA_FILENAME. (integer) " -->
<entry id="SSTICE_YEAR_START" value="0" />

<!--"The last year of data to use from SSTICE_DATA_FILENAME. (integer) " -->
<entry id="SSTICE_YEAR_END" value="0" />

Do you think, that might be an issue? Do you feel anything else might have caused the high imbalance?

 

[dbailey]

This is much more of a science question. The tunings used in the F compset will not necessarily work in the E compset. I will let @hannay chime in here.

 

[Sreerag]

Hey,
In the forcing file, which are in the server https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/ocn/docn7/SOM/ have a temperature field. I believe the SOM calculate the T_mix its own. Then what is this Temperature field means? Did this temperature uses during simulation?

Thanks in advance
Sreerag