Adjust sea ice model constants to align with ClimaSeaIce.jl#main

[taimoorsohail]

Currently, NumericalEarth.jl#main and ClimaSeaIce.jl#main cannot be run together due to a mismatch in constant naming and function definitions in src/EarthSystemModels/InterfaceComputations.

I have attempted to change the relevant files so that OceanSeaIceModel() again works with the #main branches of both projects. This was the minimal fix I could find to ensure the code ran -- not necessarily an exhaustive fix for existing bugs in the #main branches of either repo.

[codecov]

Codecov Report

❌ Patch coverage is 0% with 21 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/SeaIces/sea_ice_simulation.jl	0.00%	20 Missing ⚠️
src/EarthSystemModels/earth_system_model.jl	0.00%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[simone-silvestri]

This should be fixed in PR #164, can you review that one? Should be rather easy a part the change to the temperature interface computation which is reformulated a bit, the rest should be trivial (split freshwater fluxes in rain and snow and accumulate snow in the sea ice fluxes).

[taimoorsohail]

Thanks @simone-silvestri -- I had a look and it looks good. However, I think there are some specific changes here that are not in #164 :

1. Swap OceanSeaIceModel() input order so it is (ocean, sea_ice; kwargs...) as before
2. Adjust the ThreeEquationState default in ComponentInterfaces to be sea_ice_ocean_heat_flux = ThreeEquationHeatFlux(eltype(exchange_grid)),
3. Add a guard so that ocean.model is not accidentally passed to sea_ice specific components:

sea_ice_model(sea_ice::Simulation) = sea_ice_model(sea_ice.model)
sea_ice_model(model::SeaIceModel) = model

sea_ice_model(model) =
    throw(ArgumentError("Expected a ClimaSeaIce.SeaIceModel or sea-ice Simulation, got $(typeof(model))"))

(3) is more descriptive in case users accidentally swap the order of ocean and sea_ice in OceanSeaIceModel

[simone-silvestri]

Ok, I ll merge that one then we can also add what implemented in this PR

[simone-silvestri]

are you sure about this change? ThreeEquationHeatFlux is specialized for a ClimaSeaIce simulation to include the internal heat flux in the balance of fluxes required by the interfacial temperature. If we use eltype(exchange_grid) we loose the internal conductive flux

[simone-silvestri]

maybe instead of having a sea_ice_model guard we can specialize these functions for

heat_capacity(sea_ice::Simulation{<:SeaIceModel})

which is the more "julia" way

[simone-silvestri]

same as above, we probably should just specialize ThreeEquationHeatFlux to dispatch on

ThreeEquationHeatFlux(sea_ice::Simulation{<:SeaIceModel}, ....)

[taimoorsohail]

Thanks @simone-silvestri -- I think you are right about the ThreeEquation fix here. I will make the changes now.

[taimoorsohail]

Actually @simone-silvestri I am closing this PR now. I think the only meaningful addition was the more verbose sea ice model error, and that can be added upstream to warn the user that the input order in OceanSeaIceModel is wrong. It doesn't need to happen in the InterfaceComputations. The choice of doing OceanSeaIceModel(sea_ice, ocean; kwargs...) vs OceanSeaIceModel(ocean, sea_ice; kwargs...) is probably better discussed in an issue anyway.

[taimoorsohail]

Closed in favour of #230