Meta: Couple arbitrary parametrisation e.g. convective, orographic

[omarjamil]

No description provided.

[jatkinson1000]

• Beres single column emulator
• WRF data-driven single-column scheme?
• knowledge for datawave collaborators to use FTorch to couple their own codes
• Others?

[jatkinson1000]

Discussion in meeting:

YS - If we can couple the convective scheme it should provide knowledge for future coupling by DW members, but we should get @phz-fluids to confirm too.

Measureable goals:

• DW members approving coupling and documentation PRs with an understanding for Meta: Documentation newCAM_emulation#32 documentation, loading net, coupling in the CAM repo.
• ICCS and DW members to run CAM coupled to NN, with output produced for analysis
• ICCS and DW members to run CAM coupled to NN trained on WRF data when it becomes available

Questions:

• Do we need an issue to track the development of the WRF scheme so that it is ready for ICCS to use?

[jatkinson1000]

Discussion at the DataWave science meeting was that Ed's group is also interested in coupling a data-driven scheme to CAM.

On 19/11/2024 @omarjamil and @jatkinson1000 met with @yqsun91 who said that development of a data-driven WRF scheme is blocked until Joan is back.

In light of this we discussed that perhaps we should pivot to working with the NYU group as the ICCS resource is time-bounded.
However this requires discussion with @phz-fluids (and perhaps Aditi). This should be discussed ASAP when Pedram next joins a meeting, and we can then look at inviting NYU to join and discuss specifics with them.

[jatkinson1000]

@phz-fluids comments that this is not blocked until Joan returns and perhaps he, @yqsun91, Hamid can discuss this week.