POP in CESM: Arithmetic exception causing floating point exception

[lehner@climate_unibe_ch]

Hello everyone

I'm running CESM in the B_1850-2000_BGC-BPRP compset in 0.9x1.25_gx1v6 resolution. After many years of running the model crashes with the following error message (from gdb):

Program terminated with signal 8, Arithmetic exception.
#0 0x0000000000d1587a in vertical_mix::init_vertical_mix () at /scratch/rosa/lehner/cesm1.0/BPRD_trans/ocn/source/vertical_mix.F90:374
374 allocate (VDC(nx_block,ny_block,km,1,nblocks_clinic), &

Increasing the time step from 23 to 30 in pop2.buildnml.csh didn't help (it did help before for similar crashes).

Has anyone experienced the same issue? Is there a solution in order to get the model pass this obstacle?

With many thanks for any help!

Flavio Lehner

 

[lehner@climate_unibe_ch]

Just a short follow-up.

Eventually, when setting the time step to values above 40, the model passed the obstacle - as far as I can tell at the moment without producing significant numerical artifacts. The time step was switched back to default value (23) after two model months.

Still would appreciate sharing of similar experiences.

Best,
FL