CLM5.0 vs. CTSM5.1 in CPLHIST mode

[brendanclark]

Hello,

I have successfully run CTSM release-clm5.0 in CPLHIST mode in the past. When I apply the same methods to ctsm5.1.dev133, I get the following error for cpl files after submitting the case:

Model datm missing file file730 = '/glade/p/univ/urtg0006/Brendan/clmcrop/b.e21.BWSSP245cmip6.f09_g17.CMIP6-SSP2-4.5-WACCM.006/cpl/hist.mon/b.e21.BWSSP245cmip6.f09_g17.CMIP6-SSP2-4.5-WACCM.006.cpl.hx.atm.1h.avrg.2061-12-31-03600.nc'

I have previously used cpl input as monthly files (still 3hr/1hr time frequency, but the input files are in the format of yyyy-mm). I created a new case with daily cpl files that match this error, and the run still fails. I get this error in the cesm.log file:

0:Abort with message NetCDF: Variable not found in file /glade/work/csgteam/build/pio/41003/pio-2.5.10/src/clib/pio_nc.c at line 1164

2:Obtained 10 stack frames.

2:/glade/scratch/brendanc/b.e21.BWSSP245cmip6.f09_g17.CMIP6-SSP2-4.5-WACCM.006.ctsm5.1.dev133_dailycpl_08112023/bld/cesm.exe() [0x43d5d62]

…….

I do not know which variable or NetCDF file this error is referring to. I have attached a document outlining everything I did for these cases. If anyone knows what this error means or has experience successfully running newer versions of ctsm in CPLHIST mode, please let me know what may be going wrong.

Thanks,
Brendan

 

[oleson]

As noted in a separate email communication:
CPLHIST mode has changed significantly because of the transition to NUOPC. Older coupler history files aren't compatible with the new format, although it might be possible to convert them, but I'm not aware of anyone that has tried that. I've only very recently completed a successful CPLHIST run using the new format.
The case is here:

/glade/work/oleson/cesm2_3_alpha14a/cime/scripts/ctsm51_cesm23a14a_ne30pg3ne30pg3mg17_CPLHIST_1850

It uses cesm2_3_alpha14a which uses ctsm5.1.dev123.

 

[brendanclark]

Thanks, Keith. An update for anyone who is interested: You can still use old coupler history files to force the newer versions of CTSM in CPLHIST mode. You just have to rename the coupler history files and the variables within them to match the new format.

 

[bowen]

Hi everyone,

I'm posting my issue here since it seems related. I am running ctsm 5.1.dev090 on Cheyenne, and it is also a CPLHIST case. I received the following error in cesm log:

Abort with message NetCDF: Variable not found in file /glade/scratch/vanderwb/hpci-stack/220310-1705-1/54409/pio-2.5.6/src/clib/pio_nc.c at line 1164.

I have finished some similar simulations earlier this year, and those successful runs use the same code. This new case uses some new coupler history files (processed from b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.102 outputs, originally saved as once/daily files), so maybe there is some error when processing the files?

However I was not able to locate this file 'pio_nc.c' (this file is purged/moved?) so could not see what was wrong there...

Any insights would be appreciated!


Thank you!
Bowen


------
case: /glade/work/bowen/ctsm5.1.dev090/cime/scripts/ctsm51d090.cplhist585.staurban_GaoOneill
logs: /glade/scratch/bowen/ctsm51d090.cplhist585.staurban_GaoOneill/run/cesm.log.4107346.chadmin1.ib0.cheyenne.ucar.edu.231108-030825
CPLHIST files: /glade/scratch/yifanc17/Bowen_backup/hist.mon/

 

[oleson]

Hi @bowen , it seems like you've set stream_fldfilename_urbantv to a mesh file, not to a file that contains the building temperature data. So it's probably complaining that it can't find tbuildmax_TBD. etc.

 

[oleson]

I should also note that due to /glade/p being eliminated at the first of the year, I've moved my urban-specific data to campaign. So files that were previously under /glade/p/cgd/tss/people/oleson are now in /glade/campaign/cgd/tss/people/oleson, which is now accessible from cheyenne.

 

[bowen]

Hi Keith, yes I seem to have pointed to the wrong files. The case works now after that is fixed.

Thanks a lot for the information!

 

[bowen]

By the way, I tried to access this directory /glade/campaign/cgd/tss/people/oleson but got permission denied error on Cheyenne login/batch, Derecho login and Casper login... So it seems I really don't have the permission?

 

[oleson]

I've changed the permissions to some of the top-level directories there. Let me know if there is a specific file or files you can't access.

 

[KeerZ]

Thanks, Keith. An update for anyone who is interested: You can still use old coupler history files to force the newer versions of CTSM in CPLHIST mode. You just have to rename the coupler history files and the variables within them to match the new format.

Hello Brendan! Thanks for this post. I found it very useful! I am trying to run CTSM5.2 in CPLHIST mode. I renamed the daily coupler history files and variable names to match the new format but my case still failed. See this thread I posted. If you have run a CPLHIST simulation in CTSM before, could you share your case with me as a reference when you have time (on the forum or via email: keer.zhang@yale.edu)?

Any comments or suggestions are appreciated! Thanks!

 

[slevis]

I think that the last question was resolved in

How to run CTSM5.2 in CPLHIST mode?

Hi all, I am trying to run a simulation in the ISSP585 CPLHIST mode (compset long name: SSP585_DATM%CPLHIST_CLM50%SP_SICE_SOCN_MOSART_SGLC_SWAV) with CTSM5.2. I noticed that the coupler history files for CTSM have new NetCDF names (e.g. **.cpl.hx.atm.3h.avrg.**.nc instead of...

bb.cgd.ucar.edu