Add code changes for the implementation of the low-resolution B

[guillaumevernieres]

Still in draft mode

Work done in this PR

Interp to low-resolution:

• variance partitioning
• preparation of the perturbations for the envar
• hybrid weights
• ...

Testing

Visual check showing the variance explained by the steric height (30 members)

left is native resolution, right is on the 1/2 deg grid.

"Science" note: The explained variance above is almost the opposite to what I would have expected, and what we saw in the previous offline ensemble, or even what we see in tendencies. Check if there's a bug in the explained variance calculation.

Automated CI tests to run in Global Workflow

• atm_jjob
• C96C48_ufs_hybatmDA
• C96C48_hybatmaerosnowDA
• C48mx500_3DVarAOWCDA
• C48mx500_hybAOWCDA
• C96C48_hybatmDA

[emcbot]

Automated GW-GDASApp Testing Results:
Machine: hera

Start: Mon Jan 13 21:38:37 UTC 2025 on hfe10
---------------------------------------------------
Build:                                 *SUCCESS*
Build: Completed at Mon Jan 13 22:32:01 UTC 2025
---------------------------------------------------
Tests: ctest -j12 -R gdasapp -E atm_jjob|C96C48_ufs_hybatmDA|C96C48_hybatmaerosnowDA|C96C48_hybatmDA
Tests:                                 *SUCCESS*
Tests: Completed at Mon Jan 13 22:55:24 UTC 2025
Tests: 100% tests passed, 0 tests failed out of 53

[shlyaeva]

Did not test, but I reviewed all the code changes, they look good to me.