[joshm]

I have been looking into homogeneity of atmospheric boundary layer profiles through an empty fetch. I noticed that nutkAtmRoughWallFunction performed slightly worse than nutkRoughWallFunction.

When looking at the code to work out what the atmospheric version was doing, I came across a definition I can't work out:

As expected, both define effective viscosity as:


For nutkRoughWallFunction, E' is defined as expected, based on roughness regime. For fully rough (most practical ABL applications), it's:

However, for the nutkAtmRoughWallFunction, it is defined as:
.

I've not seen this definition in the literature, and can't work out why it's used. It means that the sand grain roughness is never defined in the equation, which seems odd.

Can anyone shed any light on this?