B grid dynamics

[simone-silvestri]

This PR will require quite some time because it is experimental and here just for documentation purposes.

[simone-silvestri]

Difference between C-grid and B-grid dynamics in the cyclone case showing the advantage of less interpolation for rheology in the B-grid:

This PR:
https://clima.github.io/ClimaSeaIceDocumentation/previews/PR55/literated/ice_advected_by_anticyclone/

C-grid (main)
https://clima.github.io/ClimaSeaIceDocumentation/dev/literated/ice_advected_by_anticyclone/

These results are expected and shown in
https://agupubs.onlinelibrary.wiley.com/doi/pdfdirect/10.1029/2021ms002523

Edit: The resolutions are different! I will produce new plots

Down here the videos of the cgrid cyclone case (main)

cgrid_anticyclone.mp4

and the bgrid case (this PR)

bgrid_anticyclone.mp4

[simone-silvestri]

Also, looking at the immersed example https://clima.github.io/ClimaSeaIceDocumentation/previews/PR55/literated/ice_advected_on_coastline/ the behavior is the expected one: given that we have impenetrable conditions on immersed boundaries, there is a small ice block that remains attached downstream the triangular element, in a physical shape. The same is not possible to obtain on c-grid without messing up quite significantly with the rheology and the momentum tendencies

[simone-silvestri]

Also noteworthy to mention, on CPUs the b-grid EVP formulation is about twice as fast as the C-grid because it avoids computing the viscosity on both cca and ffa points.

[glwagner]

How do you select B- vs C-grid dynamics? Can you give an example of the user interface + abstractions implemented?

[simone-silvestri]

The user interface would not change, only the numerical discretization of the rheology, so the user does not necessarily "see" this change, but has to keep it into account (the u and v are not on edges anymore but on nodes).

I think it would be quite messy to allow both B and C grid (it effectively duplicates a lot of the code) so my idea was just stick with one discretization.

[codecov]

Codecov Report

Attention: Patch coverage is 80.83333% with 23 lines in your changes missing coverage. Please review.

Project coverage is 74.00%. Comparing base (a922f71) to head (9c6c6a0).
Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
src/sea_ice_advection.jl	50.00%	11 Missing ⚠️
...SeaIceMomentumEquations/sea_ice_external_stress.jl	0.00%	6 Missing ⚠️
src/Rheologies/viscous_rheology.jl	60.00%	4 Missing ⚠️
src/Rheologies/elasto_visco_plastic_rheology.jl	95.65%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #55      +/-   ##
==========================================
+ Coverage   70.44%   74.00%   +3.55%     
==========================================
  Files          27       27              
  Lines         829      777      -52     
==========================================
- Hits          584      575       -9     
+ Misses        245      202      -43     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.