AD spin-up using 30y atm data - NH4 uptake error

[wvsi3w]

Dear scientists,
I have found this error when I did an AD spin-up with my cesm2.1.3 (CLM5):
problem with limitations on nh4 uptake 0.000000000000000E+000
-Infinity
ENDRUN:
ERROR: too much NH4 uptake predicted by FUN

I have read the forum and mostly there are some unresolved issues about this error. Some said it might be because of the calandar (no leap, etc) or because you set the start date to 0001, but I tried it with start date being as is (1850...) and it failed with the same error.

I haven't experienced this error when I used only one or two years of forcing data, this time I am getting it because I set the forcing to be 30 years. Is that related? if so, then how could we get passed through it if we really want to use 30y forcing (from 1901 to 1930 GSWP3)?

 

[slevis]

A way to narrow down the source of the error may be to start from a version that you haven't changed, because I would not expect that one to give an error. Then you can introduce your changes one at-a-time until you figure out which one causes the error.

 

[wvsi3w]

A way to narrow down the source of the error may be to start from a version that you haven't changed, because I would not expect that one to give an error. Then you can introduce your changes one at-a-time until you figure out which one causes the error.

Dear Sam, thank you for your response.
I did that before.
The only difference between my current case and my previous case that worked fine is that my current case uses 30 years of forcing (from the GSWP3 dataset 1901-1930), and my previous case I used only one year of forcing, 1901-1902, in spin-up.

That's it!

P.S. The reason I wanted to use 30y forcing is to have the effect of climatology on the spin-up phase because I have found that in my final simulation, if I stick to less than 30y of forcing period in spin-up, the results would be unacceptable.

 

[wvsi3w]

Also, I have chosen 23sl_3.5m as my soil layer structure.

So do you think using this soil layer (I did not use my own modified version of it, I used the default 23sl_3.5m) would create this error:
ERROR: too much NH4 uptake predicted by FUN
???

If so, what do you think we should do in these situations?
I have read this thread where you mentioned another fortran file to check:

Transient run with error message in NH4 uptake

What version of the code are you using? ctsm5.1.dev159 Have you made any changes to files in the source tree? Yes, I made changes in the code, and I am running with C13 on and with transient land use. Describe every step you took leading up to the problem: I have made changes to my code, and...

bb.cgd.ucar.edu

do you think it applies to my case too?

P.S. I tried with 20 years of forcing and it failed too

 

[wvsi3w]

--res f09_g17 --compset IHistClm50BgcCrop

./xmlchange CLM_ACCELERATED_SPINUP="on",STOP_OPTION=nyears,STOP_N=5,DATM_CLMNCEP_YR_END=1930,JOB_WALLCLOCK_TIME=03:59:00,RESUBMIT=89, RUN_STARTDATE=1850-01-01

and:

./xmlchange CLM_ACCELERATED_SPINUP="on",STOP_OPTION=nyears,STOP_N=5,DATM_CLMNCEP_YR_END=1920,JOB_WALLCLOCK_TIME=03:59:00,RESUBMIT=89, RUN_STARTDATE=0001-01-01

Either way, it fails with that NH4 error

 

[slevis]

Based on what you wrote, I see two possibilities:
- A problem in one of the GSWP3 files after 1902 and maybe you can isolate the problem based on when the simulation crashes OR
- A problem when using the 23sl_3.5m soil layer structure

The first seems unlikely because we have used the GSWP3 dataset for several years. Unless you're not on an NCAR computer, in which case maybe a file got corrupted while transferring to the computer where you're running.

The second also seems unlikely, especially if you have been using the same soil structure all along. But if this is not the default soil structure, then maybe there's a problem with it that we didn't know about. I would expect no problems with the default soil structure because we have been using that for several years.

The other thread that you mentioned may have been getting "-Infinity" from a restart file. However, the user did not update the thread to explain how they resolved the issue. I will post there to see if that user has insights.