Clarify registry descriptions for avgSurface* variables

[cbegeman]

Surface values sent to the coupler are not actually extrapolated from zMid to zTop as described. This PR changes the description of these variables in the registry to make that more clear.

[cbegeman]

@alicebarthel Do you want to give this a quick look to just confirm that my understanding of the code is correct? The only place I see these variables being assigned is here:

https://github.com/E3SM-Project/E3SM/blob/36f2208586ad996f01d14c0d16d74720d5b5bc6f/components/mpas-ocean/src/shared/mpas_ocn_diagnostics.F#L423

The surface velocity is populated in time integration schemes and also doesn't include extrapolation.

[cbegeman]

Thanks, @erinethomas, for bringing this to my attention.

[alicebarthel]

@cbegeman I agree with your assessment.
I think this code block and documentation was "forward-looking" rather than the actual implementation. Is it worth modifying the in-line comment to clarify?

[alicebarthel]

My current understanding: tracersSurfaceValue is the top-layer value, while tracersSurfaceLayerValue does the vertical averaging; normalVelocitySurfaceLayer is based on minLevelEdgeBot(iEdge) but the avgSurfaceVelocityZon/Mer are cell-centered and thus use minLevelCell. I agree with the new documentation.
@cbegeman Could you point me to the normal --> zonal velocity change code? I suppose it is layer-based and thus uses the zero velocities if/when minLevelCell != minLevelEdgeBot(iEdge) but it'd be nice to confirm.

[cbegeman]

@cbegeman Could you point me to the normal --> zonal velocity change code? I suppose it is layer-based and thus uses the zero velocities if/when minLevelCell != minLevelEdgeBot(iEdge) but it'd be nice to confirm.

@alicebarthel Yes, my interpretation of the reconstruction is that it is layer-based and uses zero velocities when minLevelCell != minLevelEdgeBot(iEdge). It doesn't have information about boundary edges or minLevelEdge* as far as I am aware.

The reconstruction code:
https://github.com/E3SM-Project/E3SM/blob/577298b47614355df9d56e307cdd87b3ea32001b/components/mpas-framework/src/operators/mpas_vector_reconstruction.F#L261-L288

And here's the place where surfaceVelocity is assigned:
https://github.com/E3SM-Project/E3SM/blob/577298b47614355df9d56e307cdd87b3ea32001b/components/mpas-ocean/src/mode_forward/mpas_ocn_time_integration_split_ab2.F#L2872-L2875

[cbegeman]

Closed in favor of E3SM-Project#7472