[CF-metadata] Dimensionless vertical coordinate values

From: Hattersley, Richard <richard.hattersley>
Date: Fri, 14 Sep 2012 10:19:18 +0100

Dear Jonathan,

Thanks for clarifying.

The original source of the confusion was example 4.3, where the
dimensionless vertical coordinate is itself one of the formula terms.
Similarly for hybrid height (where the "dimensionless" vertical
coordinate isn't even dimensionless!)

I've been through this discussion with several people prior to bringing
it to this mailing list, and no one could provide a definitive answer on
the expected behaviour. But slowly we settled on the definition that you
have confirmed. So I think CF would benefit from a couple of
clarifications. Firstly in the written description in section 4.3.2
(dimensionless vertical coordinate), and the choice of example. And also
in appendix D, where the description of what constitutes the
dimensionless vertical coordinate is not consistent across schemes -
sometimes it's "hidden" inline within the term descriptions; other times
it's a distinct sentence of its own.

Should I create a trac ticket containing a proposed clarification?

Richard Hattersley AVD Iris Technical Lead
Met Office FitzRoy Road Exeter Devon EX1 3PB United Kingdom
Tel: +44 (0)1392 885702 Fax: +44 (0)1392 885681
Email: richard.hattersley at metoffice.gov.uk Website:
www.metoffice.gov.uk
 

> -----Original Message-----
> From: Jonathan Gregory [mailto:j.m.gregory at reading.ac.uk]
> Sent: 10 September 2012 09:42
> To: Hattersley, Richard
> Cc: cf-metadata at cgd.ucar.edu
> Subject: [CF-metadata] Dimensionless vertical coordinate values
>
> Dear Richard
>
> On Mon, Aug 20, 2012 at 05:05:52PM +0100, Hattersley, Richard wrote:
> >
> > I'm glad you think the first example makes sense - it's the one that
> > makes most sense to me too! But I'm wondering if one *must* store
> > ap(z)/p0+b(z) (or a(k)+b(k) if that's the parameterisation
> in use), or
> > if one could store something else and still be a valid CF file?
>
> Yes, I think one *must* store ap(z)/p0+b(z) or a(k)+b(k) in z
> because that's
> what the standard_name and units of the vertical coordinate
> indicate z to be.
> In your example
>
> float z(z) ;
> z:standard_name =
> "atmosphere_hybrid_sigma_pressure_coordinate" ;
> z:units = "1" ;
> z:formula_terms = "ap: delta b: sigma ps:
> surface_pressure" ;
> z:positive = "down" ;
>
> z says it's the atmosphere_hybrid_sigma_pressure_coordinate.
> I think that
> ap is the pressure part of the
> atmosphere_hybrid_sigma_pressure_coordinate and
> b is the sigma part of the
> atmosphere_hybrid_sigma_pressure_coordinate. Neither
> of them alone is the
> atmosphere_hybrid_sigma_pressure_coordinate, so the
> standard_name would be wrong if you stored either of the
> components in z.
> Also, if you stored the pressure part (ap) in z, the units
> would be wrong as
> well; the units say z is dimensionless, but ap is in Pa.
>
> If this is right, should we clarify the convention text is some way?
>
> Best wishes
>
> Jonathan
>
Received on Fri Sep 14 2012 - 03:19:18 BST