Welcome to the new DiscussCESM forum!
We are still working on the website migration, so you may experience downtime during this process.

Existing users, please reset your password before logging in here: https://xenforo.cgd.ucar.edu/cesm/index.php?lost-password/

ERROR: carbon or nitrogen state critically negative

johanna_teresa

Johanna Malle
New Member
Hi,
I am running CLM5 on a local linux machine, and have done so successfully with SP mode. I am running it on a custom grid, for which I have created my own surface data/domain files and forcing data. However, recently I attempted to set up a BGC case, and keep running into the problem of carbon patch turning negative, beyond the critical cnegcrit value after ~600 timesteps (using 3hr time-steps) - see the attached lnd.log. I have tried switching to the CLM4.5 cnegcrit limits (600 instead of 60), but it only takes a bit longer before it crashes. I have also tried using different forcing datasets (all of which work with SP mode) which does not appear to be the problem. It also looked at the development of leafc over time, which starts out normal and decreases pretty much linearly until it hits the critical value. Do you have any idea what could be the problem/what I am doing wrong? When looking into this I found this issue: Clm45BgcCrop case dies due to negative leafn · Issue #66 · ESCOMP/CTSM, but this was fixed so I don't think it is the problem. Any help with this would be appreciated.
Thanks a lot,
Johanna
 

Attachments

  • config_files.zip
    22.7 KB · Views: 0
  • lnd.log.zip
    52.8 KB · Views: 1
  • version_info.txt
    8.2 KB · Views: 1

oleson

Keith Oleson
CSEG and Liaisons
Staff member
I see this in your lnd log:

dtime = 10800

which I think implies that you are running the model at 3-hourly time steps? I don't have any experience with running the model at time steps longer than 1 hour. There are a number of phenology triggers in the crop model code and probably elsewhere that may not work correctly with longer time steps.
We had a similar error in negative leaf carbon caused by this issue:


In that case a twenty minute time step was used which exposed a bug in the phenology code. I think this bug was fixed in a tag later than the one you are using so you could try implementing the bug fix in your code, but running at a 3-hourly time step in BGC mode may still not work well.
 

johanna_teresa

Johanna Malle
New Member
Thanks for the reply! Yes, I was running the model at 3-hourly time steps, but have now also tried it with 1-hourly time steps. The model runs for longer but I get the same error eventually. Weirdly it happens around the same date/time, which now makes me think it could be a problem with some input data which was somehow no problematic in the SP run. I will try implementing the bug fix in my code, thanks for the suggestion!
 

oleson

Keith Oleson
CSEG and Liaisons
Staff member
One other thing you could try is a cold start (finidat = ' '). Maybe there is a problem on initialization.
 

johanna_teresa

Johanna Malle
New Member
I implemented the suggested bug fix which in combination with running at 1hour time steps fixed my problem - the model has now ran for >50 years successfully. Thanks again!
 
Top