CESM2.2.2 BHIST with CESM2.1 CMIP6 restart files

[mark@atmos_colostate_edu]

I'm trying to initialize a CESM2.2.2 BHIST simulation with a CESM2.1 CMIP6 restart tarball, but the model crashes at initialization and the problem appears to be in the physics buffer. Are CESM2.1 atm restart files not compatible with CESM2.2.2?

Thanks,
Mark Branson

run directory: /glade/derecho/scratch/branson/BHIST_ctrl/run

 

[jedwards]

They are not. If you want to run a CMIP6 simulation you need cesm 2.1.x

 

[mark@atmos_colostate_edu]

Thanks, Jim. So I've backtracked to cesm2.1.5, set up the same BHIST case, and am trying a hybrid run using some 1980 CMIP6 restart files. It seems to give an error regardless if I use use_interp_init or not in user_nl_clm.

This is being run on Derecho.

The case directory is: /glade/u/home/branson/cesm2.1.5/cime/scripts/BHIST_ctrl
And the run directory is: /glade/derecho/scratch/branson/BHIST_ctrl/run

 

[nanr]

Hi Mark ~ Jim is off today but I might be able to help. I checked your case and see that your STARTDATE is inconsistent with your REFDATE.
Is this intentional?


1004 derecho2:/glade/u/home/branson/cesm2.1.5/cime/scripts/BHIST_ctrl > ./xmlquery RUN_REFCASE
RUN_REFCASE: b.e21.BHIST.f09_g17.CMIP6-historical.001

1005 derecho2:/glade/u/home/branson/cesm2.1.5/cime/scripts/BHIST_ctrl > ./xmlquery RUN_REFDATE
RUN_REFDATE: 1980-01-01

1006 derecho2:/glade/u/home/branson/cesm2.1.5/cime/scripts/BHIST_ctrl > ./xmlquery RUN_STARTDATE
RUN_STARTDATE: 1890-01-01

 

[mark@atmos_colostate_edu]

Hi Nan. Good catch! Thanks. That was definitely a typo. I corrected it and resubmitted the job, but unfortunately it looks like I'm getting the same CLM error as before.

 

[nanr]

Hi Mark ~

I recommend that you start fresh with this case. My concern is that you may have a legacy mismatch with your land interpolation from the bad STARTDATE.

Create a whole new case/run directory and see what happens. (You can move the old ones out of the way if you want to make any comparisons).

Nan