Why CTSM terminated after running for only one year?

[LiwenWu]

Dear Scientists,
I was running a regional SP spin-up case using my own atmospheric forcing data, land use , LAI and soil data with version CTSM-ctsm5.2.005. I submit the task and it ran with no error occurred. But the model automatically terminated the process after running and outputting data for one year, and all generated files were deleted (in the cesm/scratch/case_name/run/ directory). I have no idea about the reason. Could you please give me some advice? I have listed all the code for running the model in the file below.

What's more, I wonder if CLM-SP compset can simulate GPP and NPP? The answer seems to be no as GPP and NPP were unrecognized in the user_nl_clm. If I want to simulate water cycle and carbon cycle, is CLM-SP unable to meet it? If so, which compset can simulate GPP and NPP? I2000Clm50Bgc or I2000Clm50Cn for spin-up? Note that I am using my own landuse data and I have surfdata only for 16 pfts (surfdata_0.01x0.01_shanxi_16pfts_Irrig_simyr2000_c240401.nc), so I don't want to open the prognostic crop compset. I would like to know if the 16 pfts surfdata can meet I2000Clm50Bgc or I2000Clm50Cn compset?

Any advice would be greatly appreciated. Thank you!

 

[oleson]

You'll need to look at your log files in your run directory to determine why your model run stopped. The CLM-SP compset cannot simulate GPP and NPP, although it can simulate FPSN (photosynthesis). If you want to simulate the carbon cycle you need a BGC compset. I2000Clm50Bgc will run with a 16pft dataset.

 

[LiwenWu]

You'll need to look at your log files in your run directory to determine why your model run stopped. The CLM-SP compset cannot simulate GPP and NPP, although it can simulate FPSN (photosynthesis). If you want to simulate the carbon cycle you need a BGC compset. I2000Clm50Bgc will run with a 16pft dataset.

Hi Oleson, Thanks for your timely reply. I looked at the logfiles but saw nothing wrong. I upload the logfiles, could you please help me check if there are any errors? I tried again and again for the spin-up case but it still end run after running for one year with no error occurred. And there is an error occured from the supercomputing platform: Network error: Software caused connection abort. And the output files were no deleted anymore. May I ask if there will be any gaps (abort operation) in the model's operation when reading the data for the next year? Because the supercomputing platform I run automatically disconnects when there are no tasks for a long time, I wonder if this is the reason?
Any advice would be greatly appreciated. Thank you!

 

[LiwenWu]

@oleson

 

[slevis]

I didn't see an error in lnd.log.240618, but I saw an error at the bottom of the lnd.log.240425 file. Did you mean to share that one or is it unrelated?

When a run stops without an error, it may mean that you ran out of wallclock time, or disk space, or something along those lines...

 

[slevis]

You could work around a wallclock limitation by outputting restart files every 6 months (or something like that) and then restarting.

 

[LiwenWu]

Thanks for your suggestions. With your kindly help, my regional CLM-SP spin-up case has successfully run for years. But I found that when generating interannual files (such as ...h1.0007-12-27-00000.nc which includes data for year 0007 and year 0008), the model always needs to run for about 4 hours, while in general, the model generates a file in about 40 minutes. Is this situation normal? Is 4 hours too long compared to 40 minutes? Any advice would be greatly appreciated. Thank you!

 

[slevis]

That seems strange, but I have no insight into a potential cause.