Wave Momentum Coupling

[erinethomas]

This PR introduces a new feature to E3SMv3: Wave Momentum coupling between WW3-MPAO-EAM. This is a "stealth" feature since wave-enabled compsets are not part of standard production runs. This PR is BFB in non-wave enabled compsets, however, in wave-enabled compsets, this PR will impact the results. This PR is working towards developing a fully coupled climate "with waves" - the "LRW" configuration. Note: This feature will be used in some of the Blue-Tip simulations.

[FCC] for wave-enabled compsets when turned on

---

This new Feature is fully documented in a Design Doc:
https://acme-climate.atlassian.net/wiki/spaces/OO/pages/4902158337/Design+document+for+Wave-Momentum+Coupling

The 30+ yr PI control run with MPAS_Analysis and E3SM_Diags is documented in the simulation page:
https://acme-climate.atlassian.net/wiki/spaces/OO/pages/4901208068/Wave+Momentum+Coupling+E3SMv3+PI+Control+Run

Note: This PR must be merged AFTER #6885

[rljacob]

For wave-enabled compsets, is this PR climate changing even with the feature is off?

[erinethomas]

The results are BFB for compsets when waves are turned off. For wave-enabled compsets, by default, this feature will be ON. This feature does change the results from the current status of wave coupling, however, the current status of wave coupling in E3SM is not complete, therefore we do not yet have an estimate of the fully-coupled-climate with waves. Therefore, I would not say this PR is "climate-changing" since we have no baseline of "climate with waves" against which to compare. This PR is working towards that goal of obtaining a "climate with waves".
Based on this logic, I have removed the CC tag (it seems to be a somewhat unique case- and I am not sure which tags apply)

[proteanplanet]

The COARE algorithm needs a thorough review, especially these lines

E3SM/share/util/shr_flux_mod.F90

Lines 528 to 529 in 9676973

	wind0 = max(sqrt((ubot(n) - us(n))**2 + (vbot(n) - vs(n))**2), 0.01_r8)
	vmag = max(seq_flux_atmocn_minwind, wind0)

which should be analogous to

E3SM/share/util/shr_flux_mod.F90

Lines 338 to 343 in 9676973

	wind0 = max(sqrt((ubot(n) - us(n))**2 + (vbot(n) - vs(n))**2), 0.01_r8)
	vmag = wind0
	if (present(ugust)) then
	vmag = sqrt(vmag**2 + ugust(n)**2)
	end if
	vmag = max(seq_flux_atmocn_minwind, vmag)

but is not.

We attempted to use COARE in v2 with significant problems. Gustiness may have been the cause, now corrected for (ocn_surface_flux_scheme .eq. 0 but not (ocn_surface_flux_scheme .eq. 1. So as part of this PR, we also need a simulation illustrating that E3SM PI is not significantly different, and conserves, with COARE switched on without waves.

In addition, it needs to be explicitly shown that there is no sea level rise with waves switched on, so the B-case linked above needs to be stopped, and a new one included with the correct land initial conditions.

[erinethomas]

@wlin7 mentioned a comparison of the flux schemes (Large and Yeager 2009 vs COARE) was done as part of the V3 release. @wlin7 - what was included in this comparison and could you point us to the results?

[proteanplanet]

There is another issue. Even when COARE is corrected to include gustiness, that flux scheme is not consistent with the flux scheme over sea ice, which is part of the reason why ocn_surface_flux_scheme .eq. 0 continues to be used. U. Arizona were going to generate a COARE-equivalent scheme for sea ice, but I'm not sure where that work is at. This is a problem specifically for the wave-sea ice interaction work that is the initial focus of the capability introduced by this PR. Is there no way of including the wave terms in the default V3 scheme?

[erinethomas]

No, there is no way to couple waves into the Large and Yeager 2009 (the default v3) scheme.

[erinethomas]

The COARE algorithm needs a thorough review, especially these lines

E3SM/share/util/shr_flux_mod.F90

Lines 528 to 529 in 9676973

	wind0 = max(sqrt((ubot(n) - us(n))**2 + (vbot(n) - vs(n))**2), 0.01_r8)
	vmag = max(seq_flux_atmocn_minwind, wind0)

which should be analogous to

E3SM/share/util/shr_flux_mod.F90

Lines 338 to 343 in 9676973

	wind0 = max(sqrt((ubot(n) - us(n))**2 + (vbot(n) - vs(n))**2), 0.01_r8)
	vmag = wind0
	if (present(ugust)) then
	vmag = sqrt(vmag**2 + ugust(n)**2)
	end if
	vmag = max(seq_flux_atmocn_minwind, vmag)

but is not.
We attempted to use COARE in v2 with significant problems. Gustiness may have been the cause, now corrected for (ocn_surface_flux_scheme .eq. 0 but not (ocn_surface_flux_scheme .eq. 1. So as part of this PR, we also need a simulation illustrating that E3SM PI is not significantly different, and conserves, with COARE switched on without waves.

In addition, it needs to be explicitly shown that there is no sea level rise with waves switched on, so the B-case linked above needs to be stopped, and a new one included with the correct land initial conditions.

The COARE algorithm already includes a gustiness factor within the cor30a subroutine:

E3SM/share/util/shr_flux_mod.F90

Lines 2683 to 2690 in c199995

	! gustiness parametrisation
	Bf=-grav/ta*usr*(tsr+.61_R8*ta*qsr)
	if (Bf .GT. 0.0_R8) then
	ug=Beta*(Bf*zi)**.333_R8
	else
	ug=.2_R8
	endif
	ut=sqrt(du*du+ug*ug)

I'm not convinced we need to add gustiness again.

[proteanplanet]

I don't think that's the correct gustiness treatment. ugust comes from the atmosphere model, whereas I think what you are pointing to is in-built.

[wlin7]

. @wlin7 - what was included in this comparison and could you point us to the results?

Hi @erinethomas , here is the link to the coupled comparison simulations for the three schemes. The simulations ran to 50 years.

The diagnostics and the various comparisons (model vs obs, model vs model, and the multi-simulation comparison view (IICE) can be browsed from this simulation page.