Slab-ocean aquaplanet fails in 4xCO2 run

[mpb20@st-andrews_ac_uk]

Hi all,I've been running CESM1.2.2 (CAM4) in a slab-ocean aquaplanet configuration. Running with pre-industrial CO2 concentrations is fine, but when I run 4xCO2 simulations the model crashes after ~8 years with the following error message: Lagrangian levels are crossingSuggest to increase NSPLTVRM.
When I tried increasing NSPLTVRM from 1 to 2 (using user_nl_cam), the model failed to run with the error:
ERROR: Bad namelist settings for FV subcycling.

Has anybody had issues running a slab-ocean aquaplanet at high temperatures? Any suggestions for how to solve this?

Cheers,

Mike

 

[lanjwj]

Hi, It seems that there is a multiple relationship between those three parameters(nsplit, nspltrac and nspltvrm).
nsplit----Number of dynamics timesteps per physics timestep. If zero, a best-estimate will be automatically calculated.
nspltrac----Number of tracer advection timesteps per physics timestep.
nspltvrm----Number of vertical re-mapping timesteps per physics timestep.
Nsplit is partitioned into nspltrac and nsplit/nspltrac, with the latter being the number of dynamics timesteps per tracer timestep, possibly rounded upward; after initialization, the code quantity nsplit is redefined to be the number of dynamics timesteps per tracer timestep.
Nspltrac is partitioned into nspltvrm and nspltrac/nspltvrm, with the latter being the number of tracer timesteps per re-mapping timestep, possibly rounded upward; after initialization, the code quantity nspltrac is redefined to be the number of tracer timesteps per re-mapping timestep.

I'm not so sure, but you could try this in your namelist(user_nl_cam):

state_debug_checks=.true.
nspltvrm=2
nspltrac=2
nsplit=4
!dtime=1800


sincerely,

lanjwj

 

[mlague]

Hi Mike,

Did you ever come up with a solution to this? I'm running into a similar issue using CAM5 and a slab ocean (regular continents, not an aquaplanet). Can't do the modification suggested by lanjwj as those aren't appropriate namelist options for CAM5. DID get it to run past where it throws the "lagrangian levels" crossing by moving to a 5 minute time-step (ATM_NCPL in env_run), but that would be too computationally expensive to do my whole run at. If I try and increase the ATM_NCPL after running it past the original hiccup point, it just crashes after running a few time steps, this time without even the helpful "lagranain levels crossing" comment. It just appears to ... stop...

Similar situation where I'm running with 600ppm of CO2.

Thanks for any insight!!!

-Marysa

 

[mlague]

(edit: oh, I didn't try modifying fv_nspltvrm, fv_nspltrac, fv_nsplit - trying that now!)