Differentiated and variable AMD model constants

[glwagner]

This PR differentiates the AMD model constants for eddy viscosity and eddy diffusivity, and permits them to be functions of x, y, z.

For simplicity we will permit just one model constant specification for all tracers. It is possible to permit different constants for different tracers, but

Also, with this PR the "Verstappen" implementation of AMD is now diverging from the "Rozema" implementation. It may be time to nuke the "Rozema" implementation.

[codecov]

Codecov Report

Merging #622 into master will increase coverage by 0.38%.
The diff coverage is 71.42%.

@@            Coverage Diff             @@
##           master     #622      +/-   ##
==========================================
+ Coverage   74.63%   75.02%   +0.38%     
==========================================
  Files         117      119       +2     
  Lines        2196     2446     +250     
==========================================
+ Hits         1639     1835     +196     
- Misses        557      611      +54

Impacted Files	Coverage Δ
src/TurbulenceClosures/turbulence_closure_utils.jl	80% <100%> (ø)	⬆️
...ions/verstappen_anisotropic_minimum_dissipation.jl	94.11% <69.23%> (-5.89%)	⬇️
src/Grids/Grids.jl	57.14% <0%> (-42.86%)	⬇️
src/AbstractOperations/AbstractOperations.jl	25% <0%> (-8.34%)	⬇️
src/Coriolis/f_plane.jl	63.33% <0%> (-6.67%)	⬇️
src/Grids/grid_utils.jl	98.7% <0%> (-1.3%)	⬇️
src/Operators/interpolation_operators.jl	59.01% <0%> (-0.51%)	⬇️
src/Fields/field.jl	78.35% <0%> (-0.34%)	⬇️
src/Coriolis/no_rotation.jl	0% <0%> (ø)	⬆️
src/Grids/vertically_stretched_cartesian_grid.jl	100% <0%> (ø)	⬆️
... and 9 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1f4465d...d25abdb. Read the comment docs.