Problem about CISM when doing a run branched from CMIP6 restart files

[ccchang3@illinois_edu]

Dear the CISM/CESM community:

I tried to do a branch run based on the CMIP6 CESM2 piControl restart files.

I start the run from year 0301-01-01, and I used CESM2.1.0 B1850 f09_g17 compset.

However, I got a problem that seems to be related to CISM time manager, see the following error message from glc.log.11421111.bw.200727-130911:


(time_manager) glc date 0001-dec-27, 0.000000E+00 sec
(time_manager) glc date 0001-dec-28, 0.000000E+00 sec
(time_manager) glc date 0001-dec-29, 0.000000E+00 sec
(time_manager) glc date 0001-dec-30, 0.000000E+00 sec
(time_manager) glc date 0001-dec-31, 0.000000E+00 sec
* FATAL ERROR : (glad_main.F90:650) Valid_inputs cannot be .false. if trying
to do a mass balance time step
*******************************************************************************
Finished logging at 2020-07-27 13:22:50.307
*******************************************************************************

In my env_run.xml, I set GLC_AVG_PERIOD="yearly".

Interesting, when I set GLC_AVG_PERIOD='glc_coupling_period', the run can be integrated. However, the CISM seems to failed in generating a restart file at the end of the mid year.

Can anyone help me with this issue?

I really appreciate.

Sincerely,
Jay

 

[katec]

Hi Jay,
GLC_AVG_PERIOD='glc_coupling_period' is needed when GLC_NCPL is less than a year. What do you have your cism coupling set to?

Could you attach your glc.log file as well? If you made any changes to the xml it would be helpful to see your CaseStatus file, and if you made changes to the user_nls, please let me know what those are as well.

Thanks,
Kate

 

[ccchang3@illinois_edu]

Hi Kate:

In my env_run.xml, "GLC_NCPL" value="1". In addition, my NCPL_BASE_PERIOD" ="day".

I attached my glc.log file as well as CaseStatus file.

I really appreciate!

Jay

 

[ccchang3@illinois_edu]

To CESM community:

I would like to provide more information abut my CESM2 (as suggested by Bill Sacks).

release-cesm2.1.0

./cime
clean sandbox, on cime5.6.11
./components/cam
clean sandbox, on cam1/release_tags/cam_cesm2_1_rel_05/components/cam
./components/cice
clean sandbox, on cice5_20181109
./components/cism
clean sandbox, on release-cesm2.0.04
./components/cism/source_cism
clean sandbox, on release-cism2.1.03
M ./components/clm
modified sandbox, on release-clm5.0.14
./components/clm/src/fates
clean sandbox, on fates_s1.8.1_a3.0.0
./components/clm/tools/PTCLM
clean sandbox, on PTCLM2_180611
./components/mosart
clean sandbox, on release-cesm2.0.03
./components/pop
clean sandbox, on pop2/release_tags/pop2_cesm2_1_rel_n01
./components/pop/externals/CVMix
clean sandbox, on v0.93-beta
./components/pop/externals/MARBL
clean sandbox, on cesm2.1-n00
./components/rtm
clean sandbox, on release-cesm2.0.02
./components/ww3
clean sandbox, on ww3/release_tags/ww3_cesm2_1_rel_01


In addition, I attached my README.case/CaseStatus files below.

Finally, as the problem related to CISM occurs after the model start running. I also post all my .log files below.

Sorry that cesm.log/cpl.log/ocn.log files are too large. So I cannot post these files.

I really appreciate anyone who can help with this issue.

Sincerely,
Jay

 

[sacks]

A few follow-ups:

(1) You are using an older version (cesm2.1.0): the latest version is cesm2.1.3. Unless you have a reason to use the older version, please update to the latest release version and see if you still get this error. CESM2.1.3 contains a number of fixes relative to CESM2.1.0, which may or may not be important for your case.

(2) From the output you provided, it looks like you have some changes in components/clm. Can you please describe what you have changed?

(3) In order for us to be able to reproduce and diagnose your error, please describe every step you took, starting with create_newcase until the error.

(4) You added some additional information in an email to me. Please put that information here and I will reply here, to keep everything in one place.

 

[katec]

Hi Jay,

I'm not sure exactly what is wrong here, but some things aren't what I would expect them to be in your logs. It looks like you are not branched from the piControl. Your logs don't indicate that CISM is reading in restart information at all, and the ice sheet seems to be starting with zero mass and zero thickness. Also, it is not communicating with the rest of CESM at all in the main loop. Usually it notes mct calls while running but yours isn't doing that for some reason. There definitely seems to be something off about your coupling time steps.

If you change xml variables, you should use the "xmlchange" command to do this. That will serve several purposes: 1, it will document every change you make in your CaseStatus file (which is why I was asking for that file). 2, it will run checks to make sure that the value you set the variable to is valid for many cases.

So, if you are just trying to get a new B1850 case to run on BlueWaters, I would suggest updating to the newest CESM version, and then make all of your changes to the xml variables using "xmlchange" in the case directory. Then send us a list of the xml changes (like the CaseStatus, with the recorded changes), the changes to CLM and any user_nl changes you've made. Or take a close look at any changes you made to coupling time steps, either in CISM or other components, and make sure those changes are appropriate for these models.

Thanks!