Small changes to tilted bottom boundary layer example

[tomchor]

This PR changes two things about the BBL example:

• The variable κ was used for both the von Karman constant and for the eddy diffusivity. Now it's just used for the von Karman constant
• When calling NonhydrostaticModel we were using ordered arguments, which could make a new user interpret that they should know the order of the arguments when calling NonhydrostaticModel. Now this call is made in a more usual way with keyword arguments, where the order doesn't matter.

[navidcy]

I don't understand what you mean in the second point with "we were using ordered arguments".

The way the example was written was such that "the order of the kwargs was not important".

Syntax

f(; x, y, z)

is the same as

f(x=x, y=y, z=z)

and the order does not matter, e.g.

julia> x=1; y=2; z=3;

julia> function f(; x=0, y=0, z=0)
           @show x
           @show y
           @show z
           return x+y+z
       end
f (generic function with 1 method)

julia> f(; x, y, z)
x = 1
y = 2
z = 3
6

julia> f(; y, z, x)
x = 1
y = 2
z = 3
6

[navidcy]

let's wait for #2922