Physics cleanup

[eclare108213]

Implement calls to icepack instead of colpkg for nonstandard configurations: ISPOL, uniform_1D, single_cell.

Set a single value for pi. There are still 3 different variables: pi, pii, seaicePi (pi and pii are parameters; seaicePi is not).

Check that emissivity=1 in coupled configurations and abort if it is not.

Deprecate cesm ponds, 0-layer thermo and zsalinity. These are fully deprecated in icepack, although the MPAS-SI icepack driver still includes some active zsalinity arguments for BGC; others are commented out in case we need them. Icepack aborts if any of them are turned on. The column driver also aborts if any of these are turned on, but the code is still available if the abort is manually removed. Not tested.

Update authorship in the icepack driver.

BFB in 3-month D cases.

[eclare108213]

@njeffery: The zsalinity changes might cause the code to fail in BGC configurations, especially in standalone cases. The build system is very confusing - not sure what's for standalone and what isn't. I'm happy to fix any problems in this PR or we can do it later.

[eclare108213]

@proteanplanet: I created a new subroutine where constant values can be checked against limits or other constraints. The only thing in there at the moment is emissivity.

[proteanplanet]

Visually inspected. BFB in D-case is a good test for these changes.

[njeffery]

@eclare108213 : stand alone running of mpas-seaice is broken because "use shr_const_mod" can only be done in coupled mode. That's why we had two directories for the column constants (cesm and cice). This was merged in a previous PR in mpas_seaice_constants.F90.

[eclare108213]

@eclare108213 : stand alone running of mpas-seaice is broken because "use shr_const_mod" can only be done in coupled mode. That's why we had two directories for the column constants (cesm and cice). This was merged in a previous PR in mpas_seaice_constants.F90.

Ah, okay, I think I can fix that using #ifdef CCSMCOUPLED. Is that the only thing broken?

[njeffery]

@eclare108213 : you'll need to wrap the constant definitions as well. Anything with SHR_CONS_... needs an alternative.

[eclare108213]

But those are already wrapped, aren't they? Am I missing something?

[njeffery]

You're right! my mistake. I'll try again

[njeffery]

Missing a ,& after the pi definition
L62 pi = 3.14159265358979323846_RKIND, &

[njeffery]

Need to define seaiceLatentHeatSublimation before seaiceLatentHeatMelting

[eclare108213]

Need to define seaiceLatentHeatSublimation before seaiceLatentHeatMelting

I think it is?

I fixed the pi continuation line, will push when you get through the compile stage. Just let me know what else to fix. thx

[njeffery]

@eclare108213 : the problem seems to be that seaiceLatentHeatMelting can't be defined in terms of seaiceLatentHeatSublimation and seaiceLatentHeatVaporization because they are not parameters.

This works fine
seaiceLatentHeatMelting = 3.34e5_RKIND

[njeffery]

@eclare108213 : one more thing at it'll compile. The Makefile in analysis_members needs
-I../icepack/columnphysics

[eclare108213]

I've made those changes. Fingers crossed.

[eclare108213]

This PR remains BFB after the recent commits:

 D3.nset05.emc.physics_cleanup2.physics_cleanup.eclare108213.anvil k000 - 
    D3.nset05.emc.physics_cleanup2.icepack-integration.eclare108213.anvil k000:

    iceAreaCell:  	BFB
    iceVolumeCell:  	BFB
    icePressure:  	BFB
    uVelocityGeo:  	BFB
    vVelocityGeo:  	BFB

    totalIceExtent:  	BFB
    totalIceVolume:  	BFB
    totalSnowVolume:  	BFB
    totalKineticEnergy:  	BFB
    averageAlbedo:  	BFB


    D3.nset05.emc.physics_cleanup2.physics_cleanup.eclare108213.anvil k001 - 
    D3.nset05.emc.physics_cleanup2.icepack-integration.eclare108213.anvil k001:

    iceAreaCell:  	BFB
    iceVolumeCell:  	BFB
    icePressure:  	BFB
    uVelocityGeo:  	BFB
    vVelocityGeo:  	BFB

    totalIceExtent:  	BFB
    totalIceVolume:  	BFB
    totalSnowVolume:  	BFB
    totalKineticEnergy:  	BFB
    averageAlbedo:  	BFB

[eclare108213]

@njeffery are you okay with this PR as it is, or do you want me to wait a little longer before merging it? I think we'll need to merge it by Monday morning at the latest, preferably sooner.

[njeffery]

@eclare108213 : I'm getting a runtime error when I run stand-alone with the 'icepack' option and standard physics+snow+snicar_ad

Fortran runtime error: Actual string length does not match the declared one for dummy argument 'warningsout' (5
12/256)

Error termination. Backtrace:
#0 0xcb4afd in __icepack_warnings_MOD_icepack_warnings_getall
▶ at /gpfs/fs1/home/ac.jeffery/E3SMv2/code/20230824/components/mpas-seaice/src/icepack/columnphysics/icep
ack_warnings.F90:96
#1 0xae2a8f in __seaice_icepack_MOD_seaice_icepack_write_warnings
▶ at /home/ac.jeffery/E3SMv2/code/20230824/components/mpas-seaice/src/shared/mpas_seaice_icepack.F:17350
#2 0xb5f444 in __seaice_icepack_MOD_seaice_init_icepack_constants
▶ at /home/ac.jeffery/E3SMv2/code/20230824/components/mpas-seaice/src/shared/mpas_seaice_icepack.F:5745
#3 0xababc0 in __seaice_initialize_MOD_seaice_init
▶ at /home/ac.jeffery/E3SMv2/code/20230824/components/mpas-seaice/src/shared/mpas_seaice_initialize.F:96
#4 0x920478 in mpas_init_block
▶ at /home/ac.jeffery/E3SMv2/code/20230824/components/mpas-seaice/src/model_forward/mpas_seaice_core.F:24
0

[eclare108213]

Hmmm, I've run into that kind of problem before. MPAS seems to have some sort of internal character definition that throws errors if you try to use some other string length. I don't know what it is or what it should be -- I ended up defining a temporary character string for (I think it was the shortwave config) sent into Icepack. Very frustrating. But I don't understand why you would be getting this error in standalone when we don't get it in coupled runs.

[njeffery]

@eclare108213 : I'll look into it. By the way, this is with eclare1082131/seaice/ice_integration not physics_cleanup though I imagine it's in both. I'm testing you're branch now.

[eclare108213]

Try this:
In subroutine seaice_icepack_write_warnings in mpas_seaice_icepack.F, change
character(len=strKINDWarnings), dimension(:), allocatable :: &
to
character(len=256), dimension(:), allocatable :: &
If that works, maybe we can define a special character string length in mpas for icepack warnings. Icepack defines char_len_long=256.

[eclare108213]

If you get errors in the subroutine above that (column_write_warnings), just delete it.

[njeffery]

@eclare108213 : That worked!

[njeffery]

Approved with the following change to icepack warnings:

• character(len=strKINDWarnings), dimension(:), allocatable :: &

• character(len=256), dimension(:), allocatable :: &
  warnings

Tested in stand-alone mode with icepack and column_package options. Changes to column are BFB and differences between icepack and column_package options remain small.

[eclare108213]

Verified that the latest change compiles, runs and remains BFB.