Dear Jonathan, Philip, Martina, et al.,
first of all: many thanks for carrying this forward! This is now
becoming especially important because there is an operational service for
atmospheric composition on the horizon which will need a good standard
definition of its output.
I have a couple of points to make in response to the recent emails:
1) origin of the new proposal:
as Martina stated in her email, the current proposal was directly based
on another proposal to WMO for a new definition of chemistry variables in
the GRIB2 format. GRIB2 is the format that will be used as standard by ECMWF
for its data products and there is another committee deciding on its
adoption. The big first step will be to see general acceptance of the
proposed structure which is to separate physical quantity from species name:
one GRIB table for each of these.
2) synergies with CF:
our hope was that we could try to harmonize the GRIB2 and CF efforts as
much as possible so that a translation between the two would become
straightforward. Various people have worked on the names that have now been
suggested and these were in part different people than on the original
"atmospheric chemistry and aerosols" proposal, where in particular
Christiane textor had a leading role. Apparently this has led to some small
inconsistencies between the existing (accepted) standard names and the new
suggestions. I hope that these can be resolved without too much ado and that
the revisions can also be carried back to the GRIB2 world if that proposal
gets through at WMO.
3) conceptual remarks:
while perhaps not each and every chemical constituent listed in the
proposal will be regularly output immediately, we believe that the present
list is a reasonable compromise to cover the needs for a couple of years to
come. Since GRIB defines specific code numbers for individual variables, we
thought it is more important to settle on a rather complete list now instead
of adding names later which might lead to a lot of confusion about specific
code numbers. This philosophy differs from CF, but I believe it is possible
to reach a compromise here. What I see as most important right now is to
create the opportunity to develop a 1:1 mapping betwen GRIB2 codes and CF
standard names, and if we fuss too much about specific constituents, this
goal will be lost.
Case in point is the definition of "lumped" species, which indeed was
carried out earlier with no final conclusion. Now is the time when we
actually need something like this and so we should reach a compromise here.
There is indeed a difference between "total" and "lumped" and the latter is
vague on purpose. In the GRIB species list we developed compound groups
(similar to the JPL rate constant tables) as HOx species, NOx species,
carbon-containing species, aerosols and "project specific variables". The
latter contain all variables that are "reasonably well defined" within the
community but too unspecific or variable to be a universal standard across
communities. It may be that this is what is meant with the common_concept
proposal, but frankly this discussion goes over my head (with the limited
time I can devote to these matters). From a practicality standpoint I would
like to suggest to include these community specific standard names (and keep
the name "lumped" rather than "total"). Since there is a definition sentence
attached to each standard name, this definition could clearly indicate for
all "lumped" species that this is a community specific definition and may
vary between applications/models/projects.
4) structure:
Perhaps the most important point in bringing together the GRIB2 and CF
proposals wa the recognition that atmospheric chemsitry and aerosol
variables require a matrix rather than a vector for their definition. Hence
the separation of physical parameter and species in the GRIB2 proposal. Of
course the matrix will contain many zeros and somewhere these must be
identified (or the non-zero elements must be identified), but for the
purpose of defining the new terms it is probably best to lead the discussion
on the parameters separately from the discussion of the species names and to
delay the specific agreement on standard_names until each of these two
points has been resolved.
With respect to the parameter names I just had a conversation with
Christiane Textor who alerted me to
(i) the guidelines page on
http://cf-pcmdi.llnl.gov/documents/cf-standard-names/guidelines
and
(ii) the Wiki on
http://wiki.esipfed.org/index.php/CF_Standard_Names_-_Future_Atmospheric_Che
mistry_and_Aerosol_Terms
which contains a number of (unfortunately outdated) tables on the topic.
I believe it would be best if somehow the information on the physical
parameter space (i.e. standard_names with an X instead of a specific species
name - just as Martina has written them in her section 1) could be brought
together again in one place so that it can be expanded more easily. What is
needed in my opinion is something in between Christiane's Wiki table (which
allows too many combinations of compartment, medium, position, etc.) and the
standard_name table
(
http://cf-pcmdi.llnl.gov/documents/cf-standard-names/9/cf-standard-name-tab
le.html) which spells out each term for all compounds to which it applies.
Thus, I like Martina's "X" terms together with an explanation and a units
string.
Best regards,
Martin Schultz
< Dr. Martin G. Schultz, ICG-2, Forschungszentrum J?lich >
< D-52425 J?lich, Germany >
< ph: +49 (0)2461 61 2831, fax: +49 (0)2461 61 8131 >
< email: m.schultz at fz-juelich.de >
< web:
http://www.fz-juelich.de/icg/icg-2/m_schultz >
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Martin Schultz (m.schultz at fz-juelich.de).vcf
Type: text/x-vcard
Size: 468 bytes
Desc: not available
URL: <
http://mailman.cgd.ucar.edu/pipermail/cf-metadata/attachments/20080925/a9deb989/attachment-0002.vcf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 5634 bytes
Desc: not available
URL: <
http://mailman.cgd.ucar.edu/pipermail/cf-metadata/attachments/20080925/a9deb989/attachment-0002.bin>
Received on Thu Sep 25 2008 - 08:19:39 BST
