[CF-metadata] standard_name for coordinate variable corresponding to complex parts (real and imaginary)

From: Jonathan Gregory <j.m.gregory>
Date: Tue, 15 Jun 2010 08:24:07 +0100

Dear Benno

> This time you have understood
Good.

> I believe my proposal is a lot more useful.

Our proposals differ in two ways:

* I suggest the standard name should not be air_pressure, but should include
a phrase like fourier_transform_of_air_pressure_wrt_time. Basically that's
because if you do a Fourier transform, it's not the same physical quantity and
doesn't have the same units as air pressure, so it needs a new standard name.

* I suggest that the imaginary and real components should be in different
data variables, distinguished by the standard name. That's just the same as
we do for spatial components of vectors and tensors. When you search a file
to find the quantity you want, you can identify the real or imaginary part
by looking at the standard name. I think that's as easy, or easier, than
finding a dimension for real/imaginary and checking the coordinates.

Cheers

Jonathan

> On Mon, Jun 14, 2010 at 1:42 PM, Jonathan Gregory
> <j.m.gregory at reading.ac.uk> wrote:
> > Dear Benno
> >
> > You mean C is either "real" or "imaginary"? I think the CF-like way to do
> > this would be to have two different data variables for it, one for the real
> > part, one for the imaginary. That's like having different standard names
> > for spatial components, as we do, rather than dimensions for components.
> > As for spatial components, where "eastward", "upward", etc. are in the
> > standard name, here you would put "real" or "imaginary" in the standard name.
> > The standard name would then be not air_pressure but something like
> > ?real|imaginary_part_of_fourier_transform_of_air_pressure_wrt_time
> > I'd suggest.
> >
> > I hope I have understood now, but maybe not! Best wishes
> >
> > Jonathan
Received on Tue Jun 15 2010 - 01:24:07 BST