[CF-metadata] Use of Standard Names and Coordinate Variables (relevant to the aerosol discussion)

From: Jonathan Gregory <j.m.gregory>
Date: Sun, 8 Oct 2006 17:07:32 +0100

Dear Christiane

> It does not make sense to use coordinate variables, if
> there are only very few used (like pm? for particle size). Maybe a
> minimum of probable values for the coordinate variables could be 10.

I think it also depends, as Bryan suggested, on whether it "feels" like a
coordinate variable. Aerosols with different particle size presumably are
more like different chemical species, rather than something with a variable
dimension. Hence you put them in the standard name.

> -standardization of the rel. humidity RH at which the particle size is
> given is desirable, and as little as possible values should be used in
> the community. Hence, the RH should be part of the standard_name.

That doesn't seem necessary to me. If this is a standard value, I don't think
it needs to be stated in the standard name. It should be part of the definition
of the quantity, like the reference pressure used for potential temperature,
and the reference power level used for quantities in dB. (However, we have
many times raised the need for recording optionally what such values are that
further define standard names, and not decided on what mechanism to use. So
far it has not become critical because in the cases encountered there is only
one standard value for the parameter, as in your case.)

> -sensitivity studies/tagged tracers
> Sensitivity studies with tagged tracers are widely used in
> model studies and it seems to me, that is is useful to define a suitable
> strategie for CF.

I agree that could be useful. The consequence of this approach is that the
values of the tag name would not be standardised, but presumably defined only
for the purposes of the project. That is a similar situation to the land cover
types, for instance, where we have defined a standard name and hence a
mechanism to facilitate intercomparison, but we have not (yet) defined what
values it can take. To be more specific than "tag", perhaps it could be called
tracer_name?

> -wavelength for aerosol optical depth
> A coordinate variable could probably be used in this case, as many
> wavelengths can be used.
Good.

Cheers

Jonathan
Received on Sun Oct 08 2006 - 10:07:32 BST