Branch run didn't work but hybrid run did for FHIST compset (ERROR: Mismatch of Input Dimension)

[atmospherics]

I am trying to do a branch run based on the f.e21.FHIST_BGC.f09_f09_mg17.CMIP6-AMIP.001 restart files (1955-01-01-00000) with CESM2.1.5.

The case was created and the model was built all successfully, but after ./case.submit I got the run aborted, with errors from the cesm.log as:

dec0290.hsn.de.hpc.ucar.edu 170: check_dim ERROR: mismatch of input dimension 62336 with expected value 62100 for variable landunit

and from the lnd.log as:

Reading restart file
f.e21.FHIST_BGC.f09_f09_mg17.CMIP6-AMIP.001.clm2.r.1955-01-01-00000.nc
Reading restart dataset
check_dim ERROR: mismatch of input dimension 62336 with expected value
62100 for variable landunit
Did you mean to set use_init_interp = .true. in user_nl_clm?
(Setting use_init_interp = .true. is needed when doing a
transient run using an initial conditions file from a non-transient run,
or a non-transient run using an initial conditions file from a transient run,
or when running a resolution or configuration that differs from the initial conditions.)
ERROR:
ERROR in /CESM2.1.5/components/clm/src/main/ncdio_pio.F90.in at line 368


I looked up on this forum and found a potential solution: adding
echo "use_init_interp = .true." >> user_nl_clm
echo "init_interp_method = 'general'" >> user_nl_clm before building the model, but this didn't work.

I couldn't find any cludes yet but when I change the run type to hybrid, the run went through and worked fine.

May anyone tell me why this happened? And how should I fix the branch run configuration to make that work?

In addition, would both branch and hybrid runs based on the same restart files give similar results if I set them up with the same RUN_REFDATE according to the original restart files (without changing any other conditions / variables)?


Thanks very much for any help and suggestions :)

 

[katec]

So in a branch run, the model runs directly from the restart files, as if you are doing a "continue_run" or full restart. It sounds like your case resolution or ocean mask may be slightly different than the resolution and mask in your restart files. This will cause an error in the branch run because it does not have the correct data to exactly restart from those files, but in a hybrid run it will work as the model is initialized from a mixture of default files and your reference case (restart) files. Yes, both the branch and hybrid runs would give similar results if you set them up with the same reference files and RUN_REFDATE. For more information on branch vs hybrid starts, read this: 5. Running a Case — CIME master documentation

 

[atmospherics]

So in a branch run, the model runs directly from the restart files, as if you are doing a "continue_run" or full restart. It sounds like your case resolution or ocean mask may be slightly different than the resolution and mask in your restart files. This will cause an error in the branch run because it does not have the correct data to exactly restart from those files, but in a hybrid run it will work as the model is initialized from a mixture of default files and your reference case (restart) files. Yes, both the branch and hybrid runs would give similar results if you set them up with the same reference files and RUN_REFDATE. For more information on branch vs hybrid starts, read this: 5. Running a Case — CIME master documentation

Hi, thank you very much for the reply. I did use the same compset as the restart files, and didn't change anything other than the SSTICE, do you know any clues that this mismatch can happen? Thanks again.

 

[katec]

Hi, the SSTICE file specifies the global ocean surface temperatures and sea ice fields. If you changed this file, then the surface of the model will have different data and be incompatible with the restart files. You cannot change surface datasets in a branch run, you must use a hybrid restart to do this.