error - BalanceCheck: Tracer water balance in running iCESM1.2

[yzLiu]

Dear all,
I ran a simple PI startup case using iCESM1.2 in our machine(xdsc) following the water isotope instructions in github, the isotopic clm4 initial file was generated by the ncl script in github, and I added " init_wiso_option = 'zero' '' in user_nl_pop2. To avoid 'Total mass conservation limit exceeded' error, I added 'wtrc_check_total_h2o = .false.' to the user_nl_cam. But it failed after I sbatch my run for about 50 minutes, there are so many "isotope negative in isotope code" and " Int Neg iso *** = ***" in cesm.log. But maybe they are not the fatal error to kill my run, and many "error - BalanceCheck: Tracer water balance" may be culpable for it. But I don't know how to fix it.

Here is the context of the error:

errh2o= -4.381894153082655E-013 -14.8451020845230
endwb= 6216.47063669695 6201.36903976591
begwb= 6216.21434639855 6216.21434639855
forc_rain= 1.988741389011759E-005 0.000000000000000E+000
forc_snow= 1.231554078924164E-004 0.000000000000000E+000
qflx_irrig= 0.000000000000000E+000 0.000000000000000E+000
qflx_floodc= 0.000000000000000E+000 0.000000000000000E+000
qflx_evap_tot= 3.278035314723898E-007 -2.178812973485129E-007
qflx_surf= 0.000000000000000E+000 0.000000000000000E+000
qflx_qrgwl= 0.000000000000000E+000 0.000000000000000E+000
qflx_drain= 3.315191396972893E-007 3.315191396972893E-007
qflx_snwcp_ice= 0.000000000000000E+000 0.000000000000000E+000
ENDWB 6216.47063669695 6201.36903976591
BEGWB 6216.21434639855 6216.21434639855
QFLX_EVAP_TOT 5.900463566503016E-004 -3.921863352273232E-004
QFLX_SURF 0.000000000000000E+000 0.000000000000000E+000
QFLX_QRGWL 0.000000000000000E+000 0.000000000000000E+000
QFLX_DRAIN 5.967344514551208E-004 5.967344514551208E-004
FORC_RAIN 3.579734500221166E-002 0.000000000000000E+000
FORC_SNOW 0.221679734206349 0.000000000000000E+000
error - BalanceCheck: Tracer water balance
snl= -3
water tracer index and time step= 2 45
-2 snoliq= 0.294799937542867 3.225579168831371E-004
-2 snoice= 3.24287006598585 2.235866985577993E-002
-1 snoliq= 0.620607897560552 0.117955715345466
-1 snoice= 6.57341220770112 0.449608098954187
0 snoliq= 0.491398453745333 0.109878260331310
0 snoice= 5.38693073298354 0.443017186793315
totliq= 344619.305654046 344619.289687415
totice= 1268845.06198834 1268845.06198834

Thanks very much

Liu

 

[slevis]

Just to confirm, did you say that your simulation keeps running, but you are curious about a way to address these errors / warnings?

 

[yzLiu]

Just to confirm, did you say that your simulation keeps running, but you are curious about a way to address these errors / warnings?

Thanks, these errors did killed my simulation, but later when I tried to use CAM5, no such errors occurred. So maybe CAM4 can not support to run iCESM1.2.

 

[jiangzhu]

Only iCAM5 is available in iCESM1. Please read the paper documenting iCESM1
https://doi.org/10.1029/2019MS001663

 

[yzLiu]

Only iCAM5 is available in iCESM1. Please read the paper documenting iCESM1
https://doi.org/10.1029/2019MS001663

Thanks, I have read this classic paper. And by the way, if I change the NTASKS_ATM to a relatively large number, will it run faster?

 

[jiangzhu]

You could try increasing tasks and threads but the throughput in simulated years per day may be come saturated at some point, depending on your machine.