Updates for w

[syha]

Description:

Major updates for w obs (for radial velocity) and a bug fix for vertical interpolation in the vertical localization in height. Zgrid is now considered at either full or half levels (vertically), and at the center or edge (horizontally), just like in the model. Some cleanups and clarifications are also made.

Fixes issue

The mpas_atm model uses a height vertical coordinate with the height defined at full levels (e.g., zgrid). Most prognostic variables (ex. theta, rho, and qv) are defined at half levels, which are calculated as the middle points between the full levels. Only w (vertical velocity) is defined at full levels. The model_mod has previously only checked "oncenters" to differentiate cell center vs edges for zgrid, but this only applies to horizontal location differences at the same half levels. Now, to support QTY_VERTICAL_VELOCITY, we added "onHalf", checking full vs. half levels, as well. Along the line, we found a bug related to the vertical interpolation in height and fixed it.

Types of changes

• [x] Bug fix (non-breaking change which fixes an issue)
• [x] New feature (non-breaking change which adds functionality)
• [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
• [ ] Documentation update

Documentation changes needed?

• [ ] My change requires a change to the documentation.
• [ ] I have updated the comments accordingly.

Tests

The new model_mod.f90 has passed the test with a 2-member ensemble and a single obs.

Checklist for release

• [ ] Merge into main
• [ ] Create release from the main branch with appropriate tag
• [ ] Delete feature-branch

Testing Datasets

• [x] Dataset needed for testing available upon request
• [ ] Dataset download instructions included
• [ ] No dataset needed

[hkershaw-brown]

Along the line, we found a bug related to the vertical interpolation in height and fixed it.

Can you describe the bug you found? What was the problem, what was the fix?

[syha]

Sure. In Lines 4689-4690 in the main branch, fdata should not be always computed from zGridFace (which is now zGridFull, following the model's definition). It was supposed to be zGridCenter because we only considered half-level cases, so far. Now we split up the case for both half and full levels.

[hkershaw-brown]

There are a couple of name changes to variables, so it looks like there is a lot of code change here

[hkershaw-brown]

Sure. In Lines 4689-4690 in the main branch, fdata should not be always computed from zGridFace (which is now zGridFull, following the model's definition). It was supposed to be zGridCenter because we only considered half-level cases, so far. Now we split up the case for both half and full levels.

To clarify this is not a bug in the main branch, because no variables were on the zGridFull (zGridFace)?

[syha]

The model variable 'zgrid' was read as zGridFace in the main branch. Then it was used to compute zGridCenter, which was used for theta, qv, etc. So the lines 4689-4690 should have used zGridCenter, not zGridFace (which is now zGridFull). It was a bug, for sure, because find_vert_indices (above the lines) found vertical indices in zGridCenter.