CISM Challenge Exercise#
In this challenge, you will set up and run a T compset forced with existing output from a B compset simulation (fully coupled climate simulation with no evolving ice sheet). The experiment will force the Greenland ice sheet (GrIS) with atmospheric forcing spanning 2015-2100 that were created from a 2 degree fully-coupled SSP5-8.5 scenario experiment.
WARNING
The forcing data you will use in this exercise has not been scientifically validated yet, or ever been looked at carefully. You are entering the world of research that we deal with daily at NCAR :).
Create a case called T_GrIS_SSP585_2015_2100 using the compset T1850G at f19_g17_gl4 resolution.
Note 1: The 2 deg grid we are using in this exercise is not scientifically validated.
Note 2: In the new version of CESM, the compset names will be different and T1850G is replaced by T1850Gg (evolving GrIS) and the grid resolution f19_g17_gl4 will be replaced by f19_g17_gris4.
Note 3: The grid resolution for this case and the forcing data are both f19_g17_gl4. For a T compset, the forcing data and experiment must be at the same grid resolution.
Step 2: Make the changes to the case XML variables
Note 1: The run length is 86 years.
Note 2: The forcing data for this case can be found here: /glade/u/home/gunterl/workshops/CESM-Tutorial/data/cpl_SSP585
Step 3: Build and run the model
Click here for hints
How to I create a case a case that is not scientifically supported?
You need to add the option
--run-unsupported
at the end of your call to ./create_newcase
How do I make changes to case xml variables?
Make your changes to xml variables using the command:
./xmlchange
How do I compile?
You can compile with the command:
qcmd -- ./case.build
Click here for the solution
Create a new case T_GrIS_SSP585_2015_2100 with the command:
cd /glade/campaign/cesm/development/cross-wg/tutorial/release-cesm2.1.5/cime/scripts
./create_newcase --case ~/cases/T_GrIS_SSP585_2015_2100 --compset T1850G --res f19_g17_gl4 --run-unsupported
Case setup:
cd ~/cases/T_GrIS_SSP585_2015_2100
./case.setup
Change the run length:
./xmlchange STOP_N=86,STOP_OPTION=nyears
If you are completing this exercise outside of the tutorial, change job queue and account number. For instance:
./xmlchange JOB_QUEUE=main,PROJECT=UESM0013
Modify the case XML variables:
./xmlchange DLND_CPLHIST_DIR=/glade/u/home/gunterl/workshops/CESM-Tutorial/data/cpl_SSP585
./xmlchange DLND_CPLHIST_CASE=b.e21.BSSP585cmip6.f19_g17.CMIP6-SSP5-8.5.001
./xmlchange DLND_CPLHIST_YR_START=2015
./xmlchange DLND_CPLHIST_YR_END=2100
./xmlchange RUN_STARTDATE=2015-01-01
./xmlchange DLND_CPLHIST_YR_ALIGN=2015
Confirm that you have set up the paths and file names correctly by running:
./preview_namelists
and examine the generated file,
CaseDocs/dlnd.streams.txt.sno.cplhist
Set up the environment variable for qcmd. If using tcsh or csh:
setenv PBS_ACCOUNT UESM0013
and for bash:
export PBS_ACCOUNT=UESM0013
Build and submit:
qcmd -- ./case.build
./case.submit
If you see errors saying the dlnd forcing files are missing or unable to be downloaded since they lives outside of the input_data_root, then you may need to copy all of the forcing files into your case run directory.
cp /glade/u/home/gunterl/workshops/CESM-Tutorial/data/cpl_SSP5/* /glade/scratch/derecho/[username]/T_GrIS_SSP585_2015_2100/run/
And then submit the run again. When the run is completed, look into the archive directory for: T_GrIS_SSP585_2015_2100.
(1) Check your archive directory on derecho (The path will be different on other machines):
cd /glade/derecho/scratch/$USER/archive/T_GrIS_SSP585_2015_2100/glc/hist
ls
(2) Take a look at the contents of a file using ncdump.
ncdump -h T_GrIS_SSP585_2015_2100.cism.h.2016-01-01-00000.nc
You will notice that CISM outputs are written yearly (by default) even though the time step is 0.1 year. Changes in ice sheet variables that are typically looked at (such as changes in ice mass or ice thickness) will be captured at this output frequency.