[CF-metadata] Hoeck/Stockhause standard names

From: Pamment, JA <J.A.Pamment>
Date: Tue, 1 May 2007 14:14:49 +0100

Dear Jonathan and Heinke

>
> > Looking at the units of these quantities we can write W s m-2 = J
m-2
> > and we can think of these accumulated fluxes as being the energy
that
> > has passed into or out of the atmosphere through a surface during
some
> > period of time. This leads me to come up with the following
> > alternative names for the quantities you describe:
> >
change_over_time_in_atmospheric_energy_content_due_to_surface_downward_l
> > atent_heat_flux; J m-2
> etc.
>
> That is true, but it has a slight drawback in that "energy_content" is
not
> precisely defined. In fact the standard name for atmos energy content
has
> help which appeals for help in defining it! Does it mean dry static
energy
> or whatever? In this case that feels like a distinction we would
rather
> not have to worry about.
>
> Another possibility, which would be consistent with the guidelines, is
to
> use the construction integral_of_X_wrt_time. There are actually two
such
> names in the table.
>
Ah, that's a good point - I had forgotten about the vagueness of the
"energy_content" definition. Clearly, that is something that needs to
be addressed separately. In the meantime, I think your suggestion of
using "integral" is a good one. That would give us the following names:
integral_of_surface_downward_latent_heat_flux_wrt_time; W s m-2
integral_of_surface_downward_sensible_heat_flux_wrt_time; W s m-2
integral_of_surface_net_downward_longwave_flux_wrt_time; W s m-2
integral_of_surface_net_downward_shortwave_flux_wrt_time; W s m-2
integral_of_toa_net_downward_shortwave_flux_wrt_time; W s m-2
integral_of_toa_outgoing_longwave_flux_wrt_time; W s m-2
These have the added advantage of being shorter than the names I
suggested.
Suitable time bounds would need to be specified to indicate the period
over which the energy flux is integrated.

>
> > I am not entirely clear as to whether this is evaporation of snow
> > lying on the surface, or of falling snow. Please could you clarify?
> > In either case "amount" is used in standard names to mean "mass per
> > unit area" which would not be appropriate for units of metres. I
> > would suggest using "depth" rather than "amount".
>
> Sorry to be awkward: although common for this kind of use, we have so
far
> avoided using "depth" with that meaning because of its other meaning
of
> downward vertical coordinate. There are existing names of this kind
which
> are probably analogous:
>
> lwe_thickness_of_surface_snow_amount
> lwe_thickness_of_water_evaporation_amount
>
> which are quantities in m.
That's fine: I'm glad you're keeping me on the straight and narrow,
Jonathan! So, rather than "depth" we could have,
lwe_thickness_of_surface_snow_sublimation_amount; m
(assuming that we are talking about lying snow).
>
> > Also, if the snow
> > is being converted directly to water vapour we should call it
> > sublimation rather than evaporation.
> I agree.
>

>
> > > > forest_area_fraction
> >
> > We don't currently have any standard names that refer to "forest".
Is
> > there a clear definition of the term, for example specifying which
> > types of vegetation are included and what density of vegetation is
> > required for an area to be classified as forest?
>
> I would advocate avoiding new X_area_fraction names in favour
> of using area_fraction itself with an appropriate coordinate variable,
as
> suggested by the help for area_fraction:
>
> "To specify which area is quantified by a variable of area_fraction,
> provide a
> coordinate variable or scalar coordinate variable of land_cover or
> surface_cover."
>
> The values of land_cover and surface_cover are not (yet) standardised.
> This
> approach will avoid a large number of area_fraction names.
>
If we do define a name of forest_area_fraction then we certainly have to
define "forest" which I appreciate may not be as straightforward as it
sounds. It is also true that adding a specific name for every possible
vegetation type would lead to a lot of names, so using a coordinate
variable would seem like the natural approach.

On the other hand, if we use area_fraction with a non-standardised
land_cover/surface_cover coordinate variable are we making the job of
comparing datasets more difficult? The concern I have with using the
non-standardised coordinate variables is that different datasets can use
"forest" to mean different things but those differences may not be very
obvious. A compromise would be to use area_fraction +
land_cover/surface_cover and standardise the values of land_cover and
surface_cover. What do people think about this?

> > I would be interested
> > to know if people have further views on whether we should define
> > standard names for cell measures and cell bounds.
>
> I think the argument is stronger for cell measures, which could
> conceivably
> be data variables in their own right, than for cell bounds, which must
> share
> the metadata of the coordinate variables they bound, so do not need
their
> own.
>
> Cheers
>
> Jonathan
>

Best wishes,
Alison

------
Alison Pamment Tel: +44 1235 778065
NCAS/British Atmospheric Data Centre Fax: +44 1235 446314
Rutherford Appleton Laboratory Email: J.A.Pamment at rl.ac.uk
Chilton, Didcot, OX11 0QX, U.K.
Received on Tue May 01 2007 - 07:14:49 BST