Feature/punyswrad

[DeniseWorthen]

1. Issue changes required for restart reproducibility in s2s #15

• Adds coszen to restart file reading/writing (note: existing CICE5 ICs contain coszen but it is not being read).

• Adds ice_aggregate call at initialization

• Adds ability to read halo points in ice restart file for cases when they don't currently exist. If restart_ext is set .true. but the ice is from a non-CICE source (ie, cfsr, cpcice), CICE will not read the extra points; however they will be written to the restart file.

• Replaces ice_shortwave routine with the equivalent routine from CESM_CICE5. The consistent difference is the conditional for determining when the sun is below the horizon. It is now set by calculating the net SW down > puny instead of cozen>puny.

• Replaces constants in the cice_cap with specified R*8 equivalents.

2. Issue timestamping of field dumps #14

• When field dumps are enabled, all fields at a given import or export time will be written to a single file which has a filename timestamped with the applicable time. This replaces the existing field dumping, which writes all time values of a single field into a separate file (one for each field imported or exported).

Figures are shown below for ice concentration and surface temperature change from the IC after 5 days for ufs-s2s-model using the current develop branch of CICE5 compared to the feature/punyswrad branch. The evolution of these fields from the IC in the feature/punyswrad branch is nearly identical to the evolution in the develop branch.

[binli2337]

The changes are good.

[DeniseWorthen]

I would like to push another change to this branch for CMEPS integration.

In CMEPS (as part of the CICE6 build) we want to remove the mediator calculation of air density and air potential temperature and have those fields calculated in the CICE cap. For CICE5 we have always calculated air potential temperature in the Cap, but calculated air density in the mediator.

The change will be to calculate air density in the Cap with the same formula and fields as used in the mediator.

Unfortunately, the air density calculation in the NEMS mediator is hard-coded. It is not conditional on the field being actually connected. To avoid having to make a change in the NEMS mediator also, I would propose that that air density still be advertised and connected but that the air density field be re-calculated in the CICE5 cap for both NEMS and CMEPS.

Proposed change:
DeniseWorthen/CICE5@feature/punyswrad...DeniseWorthen:feature/calc_rhoa