Lateral diffusion with VVL in Z coordinates - For Discussion

[timgraham]

When using S coordinates in NEMO ln_traldf_hor=True implies true horizontal mixing (and uses uslp & vslp), ln_traldf_level implies mixing along model levels and ln_traldf_iso implies mixing along isoneutral surfaces.

In Z coordinates ln_traldf_hor and ln_traldf_level have previously been considered as the same thing (mixing along model levels) but Rachel has raised the point that this may not be correct in the VVL case where adjacent model grid cells could be offset from each other in the vertical (particularly with tides).

Should the behaviour be changed so that in Z coordinates with VVL ln_traldf_hor is true horizontal mixing and ln_traldf_level is mixing along model levels? Perhaps this will happen anyway as part of the simplification of vertical coordinates?

[timgraham]

This was opened for discussion but no one ever discussed it so I'll close it.