NOCOMP model with accelerated decomposition

[rosiealice]

Purpose:

---

Description:

---

Compset long name
1850_DATM%QIA_CLM60%FATES-SP_SICE_SOCN_MOSART_SGLC_SWAV

---

Model grid long name
ne30pg3_tn14

---

Case directory:

• Locally on betzy (if still available):
  /projects/NS9560K/noresm3/cases/$user/$CASE

• On github:
  https://github.com/NorESMhub/noresm_simulations/tree/$CASE

---

Code version on github (either NorESMhub or user's fork):

• https://github.com/NorESMhub/noresm/tree/<code_version>
• https://github.com/username/noresm/tree/<code_version>

---

Diagnostics:

• Diags (if available):
  As examples:
• https://ns2345k.web.sigma2.no/datalake/diagnostics/ADF/plots/<$CASE_vs_$BASELINE>/website/index.html
• https://ns2345k.web.sigma2.no/datalake/diagnostics/ADF/plots/<$CASE_vs_$OBS>/website/index.html

---

Output:

• Raw output on nird (if still available):
• CAM Climos:

---

Contacts:

---

Extra details:

[kjetilaas]

I tested this, and got the following error message:

Using cold start initial conditions
hist_htapes_build Initializing clm2 history files

htapes_fieldlist ERROR: GPP in fincl( 4 ) for history tape
1 not found
ENDRUN:
ERROR: ERROR in histFileMod.F90 at line 887

[kjetilaas]

Here is the link to the setup (using commit df05388):
https://github.com/kjetilaas/Run_NorESM_script

Only difference from description above is that compset is SROF instead of MOSART.

Link to case on betzy (temporary):
/cluster/work/users/kjetisaa/noresm/n1850.FATES-NOCOMP-AD.ne30_tn14.alpha08d.20250102/run

The same setup (script) works without accelerated decomp.
And the same setup (script) works with CLM (not FATES) and AD.

[rosiealice]

Looks like it is the GPP variable (which is a CLM variable) in the fincl which is causing the issue. I wonder if this is turned on by default when we turn on AD? Is it present in tbe user_nl_clm ?

[kjetilaas]

user_nl_clm only includes:

fates_paramfile = '/cluster/work/users/rosief/git/calibration_parameter_files/SP_calib/fates_params_det_alpha09.nc'
use_fates_nocomp=.true.
use_fates_fixed_biogeog=.true.

[rosiealice]

Right. So it must be being set automatically... Is there an fincl in the lnd_in file in the run directory?

[kjetilaas]

Yes, it is. Adding this line to the user_nl_clm solved the problem:
hist_fincl1='TOTSOMC','TOTSOMN','TLAI','FATES_GPP','FATES_NPP','TWS'

[rosiealice]

Oh cool. That's good news...