sorry if I was not clear enough with my question.
I have model outputs which are not equally spaced in vertical and I'd like to know how to calculate for instance weighted mean in vertical temperature.
The average Ferret computes with the @AVE transformation is already weighted by the thickness of each layer.
The only caveat is that the weighting is correct only when your dataset provides the edges of the gridboxes. For example, suppose this is how your temperature data looks like:
+--- T(1) --+------- T(2) ----------+ . . . . .
z(1) z(2)
Ferret cannot know what weights it should give to T(1), T(2), etc. even if it knows the depths z(1), z(2), etc. at which the temperature values are defined.
You need to tell Ferret the depths of the "+" symbols in the diagram above. If you use a netCDF file, it's achieved by the "_bnds" coordinates. See the CF convention.
If you define your axes for yourself, you use the EDGES qualifier.
Please note that I've written the above without consulting the Ferret manual and so details may be wrong, but this is basically how it works, I think.