ice: Vertical thermo error

[agustin@unex_es]

I am running CESM2.1.1 with --compset BSSP585cmip6 --res f09_g17. I has been running it from year 2015 up to year 2051. I have had this problem before and I have solved it  by increasing the  parameter dt_count in the user_nl_pop namelist from 24 to 48. But now it has ocurred again with the parameter dt_count set at 48.   Any help will be wellcome. Thanks,  AgustinzTsn:  -100.019816725937
 Tmin:  -100.000000000000
 Tmax:  2.928841054192124E-004
 zqsn:  -180004755.928577
 zap_snow_temperature: temperature out of bounds!
 istep1, my_task, i, j, k:    30099110          44           8          13
           2
 zTsn:  -100.018929660995
 Tmin:  -100.000000000000
 Tmax:  8.585876909756977E-004
 zqsn:  -180004136.137100
  max rss=2560.5 MB
 memory_write: model date =   20511227       0 memory =      -0.00 MB (highwater)       2560.54 MB (usage)  (pe=  320 comps= OCN)
  max rss=1040.5 MB
  max rss=474.3 MB
 memory_write: model date =   20511227       0 memory =      -0.00 MB (highwater)        471.74 MB (usage)  (pe=  160 comps= cpl ATM ICE GLC WAV)

 Starting thermo T < Tmin, layer           1
 zTin =  -100.317182657607
 Tmin =  -100.000000000000
 istep1, my_task, i, j:    30099132          44           8          13
 istep1, my_task, iblk =    30099132          44           2
 category n =            2
 Global block:         903
 Global i and j:          67         348
 Lat, Lon:   65.1595262563077        23.6830892204639
 aice:  0.999996146423672
 n:            2 aicen:   7.442815936059255E-003
 ice: Vertical thermo error
 ERROR: ice: Vertical thermo error
Image              PC                Routine            Line        Source
cesm.exe           0000000003135BEE  Unknown               Unknown  Unknown
cesm.exe           0000000002DCCE72  shr_abort_mod_mp_         114  shr_abort_mod.F90
cesm.exe           000000000177BC84  ice_exit_mp_abort          46  ice_exit.F90
cesm.exe           0000000001AEF34A  ice_step_mod_mp_s         569  ice_step_mod.F90
cesm.exe           00000000018C7C64  cice_runmod_mp_ci         186  CICE_RunMod.F90
cesm.exe           000000000176D3FA  ice_comp_mct_mp_i         563  ice_comp_mct.F90

 

[mmills]

My method for getting through these vertical thermo errors is to make a small pertubation in one of the atmosphere parameters: clubb_gamma_coef = 0.308 => 0.30800001 Apparently this is just a random crash that can be avoided by introducing a small pertubation, without affecting the climate. 

 

[agustin@unex_es]

Hi, everybod, the method proposed by Mike Mills seems to work. It has been running for 20 years whitout crashing.

Thanks Mike

Agustin

 

[dbailey]

This still concerns me. The message says that the starting guess for the snow temperature is less than -100C. This is physically unrealistic. I believe something else in the system is going wrong. This is very different than the intermittent flux conservation errors. I suppose this was fixed by the perturbation to club_gamma_coeff, but if this happens again, you should output every step coupler output and look carefully at all of your fields.