CICE case with BL99 scheme init_vertical_profile error

[wangxiong]

Hi, guys. Since I am a rookie in CICE, so in recent few days, I thought why not just stick to the forcing data downloaded from Github and try to run several simple cases with different settings. In this specific case, I use the gx3 grid and change nothing but switch the thermodynamic scheme from "mushy layer" to "BL99", and all the changes in ice_in I have done are according to CICE/icepack/configuration/scripts/options/set_nml.thermo1. But after I submitted the task , I got the error message. I put the brief error message down below:

Finished writing ./history/iceh_ic.1998-01-01-00000.nc

(init_vertical_profile)Starting zSin < min_salin, layer 1
(init_vertical_profile)zSin = 5.467137390670132E-002
(init_vertical_profile)min_salin = 0.100000000000000
(icepack_warnings_setabort) T :file icepack_therm_vertical.F90 :line
868
(init_vertical_profile) init_vertical_profile: Starting zSin < min_salin, layer
(icepack_warnings_aborted) ... (thermo_vertical)
(icepack_warnings_aborted) ... (icepack_step_therm1)
(icepack_step_therm1) ice: Vertical thermo error:
(icepack_warnings_aborted) ... (icepack_step_therm1)
(icepack_warnings_aborted) ... (icepack_step_therm1)
(icepack_warnings_aborted) ... (icepack_step_therm1)
(icepack_warnings_aborted) ... (icepack_step_therm1)

(abort_ice)ABORTED:
(abort_ice) called from ice_step_mod.F90
(abort_ice) line number 380
(abort_ice) error = (step_therm1)
application called MPI_Abort(MPI_COMM_WORLD, 31) - process 2

Acoording to error message, I thought it is clearly unreasonable that zSin is smaller than a minimum value and since BL99 is a scheme with fixed salinity profile, so zSin might be defined somewhere in the program (which I am not sure) and it might not be a good choice trying to change the zSin. So I located the min_salin in icepack_parameters.F90 and change the value from 0.1_dbl_kind to 0.001_dbl_kind. After that I re-compiled the code and this time the model can run successfully.
But still, I think change the default value preset in the code isn't the proper way to tackle this problem.

So Is there any way to avoid this error or deal with it more properly?

 

[eclare@lanl_gov]

Hi Wangxiong,
I'm glad to hear that you were able to get the code to run by changing the minimum value. I'm not sure how safe that is to do, though. That minimum probably exists in order to keep the numerical values away from potential instabilities. In general, when changing parameter settings that fundamentally change assumptions in the model (like whether the salinity profile is fixed or prognosed, or changing the number of layers or thickness categories), the model needs to be re-spun-up from no-ice initial conditions, in order for the heat, salt, etc content to be internally consistent.
Elizabeth

 

[wangxiong]

Hi Elizabeth,
Thank for your reply. I am pretty unfamiliar with the concept of spun up. If not too much bother, can you tell me a bit more about how to carry out the spun up procedure, or recommend some paper or material about that is also good enough. Thank you.
wangxiong