How do I use B compset history output to create SST/ICE data files to drive an F compset?

[hannay]

From user:<br /><br />I've seen in the forum that you know a lot about software tools  
related to CESM. I did however not find a solution or tool which  
explains me how the input data fields for CAM are done (I will use  
CESM1.0.5 with finite volume f90_f09). I can download some input  
fields by default from the svn repository but I need to create input  
fields by myself using SST/ICEFRAC from a previous simulation.

Where could I find this information? I'm sure that I'm not the first one...

 

[hannay]

The icesst tool can be found at:ftp://ftp.cgd.ucar.edu/archive/icesst/

 

[fsche@mech_kth_se]

Thank you so much for providing the link to icesst tools. That was somehow the dead end for me sor far... Is there a reason why this is not provided anymore by default for CESM1?I will try to follow the description using monthly mean TraCE output files (so SST replaced by TS and aice by ICEFRAC) after interpolating TraCE fields (T31) first onto my target grid (f09_f09).I'll let you know how it works.

 

[hannay]

The tool is not supported because of lack of resources. This is why it is not part of the release.
But this is the code what we use to create SST-sea ice dataset and it is pretty straightforward to use. 

 

[zwang19@masonlive_gmu_edu]

Hi, I am trying to use the icesst tool provided here.
My question is, regarding the sample command provided in README:

./regrid -i MODEL.had+oiv2.ice.mnly.49-01.unf.nc
-s MODEL.had+oiv2.sst.mnly.49-01.unf.nc
-g /fs/cgd/csm/inputdata/atm/cam/coords/gauss_8x16.nc
-o sstice.nc

I am curious about the –i ice data input: should it be aice from CESM ice component or ICEFRAC from the atm component?

Besides, I am running an E_1850C5 compset with f45_g37 resolution:

ATM_GRID = 4x5 NX_ATM=72 NY_ATM=46
LND_GRID = 4x5 NX_LND=72 NX_LND=46
ICE_GRID = gx3v7 NX_ICE=100 NX_ICE=116
OCN_GRID = gx3v7 NX_OCN=100 NX_OCN=116

If I follow the userguide “CAM: How do I use B compset history output to create SST/ICE data files to drive an F compset?” the aice and SST are of different resolutions, so I must interpolate one of them to put into one nc file? What should I do then if my F compset is also of f45_g37 resolution? Do I also have to change SSTICE_GRID_FILENAME to domain.ocn.gx3v7.090903.nc?

Thank you,
Zaiyu

 

[fsche@mech_kth_se]

I did not succeed to convert the TraCE output of monthly mean TS (used as SST) and ICEFRAC (instead of aice). I made a test with the existing file  sst_HadOIBl_bc_0.9x1.25_clim_c040926.nc which should have the desired information to run with regrid. This did not work either.I wonder if somebody could help me trying to convert the TraCE output to input fields for CAM. I guess "ncks -v ice_cov ice_cov.nc sst_cpl.nc" should be "ncks -A ice_cov.nc sst_cpl.nc" but I mainly failed with the next step. I downloaded the files from the earthsystemgrid:trace.10.14900-14351BP.cam2.h0.TS.0710101-0765012.nc to be used as sst_cpl.nctrace.10.14900-14351BP.cam2.h0.ICEFRAC.0710101-0765012.nc to be used as ice_cov.ncCheers,Frederik

 

[zwang19@masonlive_gmu_edu]

I think my problem is that the ATM_GRID is FV grid while ICE_GRID is gx3v7 (Greenland pole v7 3-deg grid).

I was running an f45_g37 resolution E_1850C5 compset (CAM, CLM, and CICE active with Slab Ocean Mode plus stub glc).
And I could only extract SST from CAM and aice from CICE (or compute aice from CAM), either way I need to re-grid SST (and ice_cov if also from CAM) into gx3v7 to force CESM1_2_0 F compset at resolution f45_g37.

But how to re-grid to gx3v7 ?

Thanks,
Zaiyu

 

[fsche@mech_kth_se]

I still didn't find the problem. Using the AMIP file sst_HadOIBl_bc_0.9x1.25_1850_2008_c100127.nc, I created the required inputfiles for bcgen:ncks -v date,datesec,SST_cpl and ice_cov etc from the AMIP file above, renamed it to SST and ICEFRAC, respectively and thenbcgen -i test_ice.nc -s test_sst.nc -g sst_HadOIBl_bc_0.9x1.25_1850_2008_c100127.nc -o test.ncyields the error:
forrtl: severe (24): end-of-file during read, unit 5, file stdinThe test-files should be exactly those required... I used gmake settingFC=ifort
FFLAGS += -OLIB_NETCDF := /software/apps/netcdf/4.2/i1214-hdf5-1.8.11-parallel/lib
INC_NETCDF := /software/apps/netcdf/4.2/i1214-hdf5-1.8.11-parallel/include

 

[zwang19@masonlive_gmu_edu]

I later adapted another way by imitating the default SSTICE_DATA_FILE:

/glade/p/cesmdata/cseg/inputdata/atm/cam/sst/sst_HadOIBl_bc_1x1_clim_pi_c101029.nc

and modify my sstice.nc file accordingly, especially to regrid to 1x1 grid resolution.
Then the F compset ran without issue! No need to re-grid to gx3v7 or reset the SSTICE_GRID_FILE.
Hope this piece of information helps,

Zaiyu

 

[fsche@mech_kth_se]

Did you do any modifications to the data prior to running bcgen for the time interpolation?

 

[hannay]

You shouldn't use the regrid/bcgen with sst_HadOIBl_bc_0.9x1.25_1850_2008_c100127.nc. This dataset is already regrid and time-diddled.

Instead please try to use:
MODEL.SST.HAD187001-198110.OI198111-201303.nc
and
MODEL.ICE.HAD187001-198110.OI198111-201303.nc

These files can be find at:
ftp://ftp.cgd.ucar.edu/archive/SSTICE/

If you want to regrid at fv_1.9x2.5.nc, the command would be:

./regrid -i MODEL.ICE.HAD187001-198110.OI198111-201303.nc
         -s    MODEL.SST.HAD187001-198110.OI198111-201303.nc 
         -g   fv_1.9x2.5.nc
         -o   sstice_fv_1.9x2.5.nc

Please give it a try.

 

[hannay]

The icesst tool will not allow you to regrid to gx3v7.
However, you don't need to regrid to gx3v7.
If you provide the SST on a standard 1x1 grid, the model will be interpolate on the fly.

 

[hannay]

See also the thread "ncl scripts to create SST/ICE data files to drive an F compset using B compset history files"https://bb.cgd.ucar.edu/ncl-scripts-create-sstice-data-files-drive-f-compset-using-b-compset-history-files 

 

[hannay]

Please check out the link:
http://www.cesm.ucar.edu/models/cesm1.2/cesm/doc/usersguide/x2288.htmlIt shows step by step how to create the SST/ICE data files to drive an F compset using files from a previous simualtion See also the thread "ncl scripts to create SST/ICE data files to drive an F compset using B compset history files"https://bb.cgd.ucar.edu/ncl-scripts-create-sstice-data-files-drive-f-compset-using-b-compset-history-files