feat: add P3 ice microphysics model

[haakon-e]

This pull request adds support for the 2-moment microphysics scheme with P3 ice microphysics ("2MP3") to the codebase, including a new test configuration, initial condition, diagnostics, and all required model infrastructure. The changes ensure that the new scheme is handled correctly throughout the precipitation, cache, initial condition, and diagnostic modules, and integrates it into CI and post-processing.

Support for 2MP3 microphysics scheme:

• Added the Microphysics2MomentP3 model throughout the codebase, including imports, cache allocation, precipitation tendencies, and diagnostics, enabling full support for the new scheme. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

New test configuration and initial condition:

• Added a new single-column test configuration file for the 2MP3 scheme (single_column_precipitation_2MP3_test.yml), and created the corresponding initial condition ColdPrecipitatingColumn and state types (PrecipState2MP3). [1] [2] [3] [4]

Continuous integration and post-processing updates:

• Integrated the new test into the CI pipeline and post-processing plotting routines, ensuring automated testing and visualization for the 2MP3 scheme. (.buildkite/pipeline.yml, post_processing/ci_plots.jlR627)

Parameter infrastructure updates:

• Extended the parameters struct to include microphysics_2mp3_params for the new scheme. [1] [2]

Diagnostics and output compatibility:

• Updated diagnostics functions to correctly handle the new scheme, ensuring all relevant output variables are available for 2MP3 runs. [1] [2] [3] [4] [5] [6] [7] [8]

[haakon-e]

@trontrytel @sajjadazimi could you take a quick look at this PR?

It's probably not quite ready to merge, but it does successfully take one time step (on my local machine).

Let me know if the overall design is reasonable enough to be merged.

In future PRs, I aspire to refactor the 2M as to remove the need for q_sno. For now, I (hope) q_sno is always zero there should be no impact carrying this variable around when in reality doing P3.

[haakon-e]

This change is part of the following stack:

• rft: conditionally load microphysics parameters #3963
  • feat: add P3 ice microphysics model #3959 ◀

_{Change managed by git-spice.}

[haakon-e]

Feedback from microphysics team: Once this is done,

• delete the 2M-only option.
• remove the distinct non-equil / equil options.

We want:

• equil, 0M
• non-equil 1M
• non-equil 2M+P3