[CF-metadata] stations and trajectories (the OTS standard)

From: NGalbraith <ngalbraith>
Date: Fri, 17 Jun 2005 11:47:47 -0400

> I think Jonathan's suggestion of extending the Standard
> Names to cover the requirements of OTS goes against the
> idea we had at GO-ESSP to avoid increasing the area of
> overlap between the two vocabularies.

This sounds like a great opportunity to develop a set of
CF standard names for subsurface parameters that will map
well to the BODC codes.
 
> What I think we need in OTS is a mechanism to link to
> either vocabulary or perferably both.

Early on in OTS we discussed allowing several naming
conventions, including EPIC, which I'd lobbied for but
have now abandoned. The problem with this is that software
*currently* available to the average user isn't going to
be capable of handling multiple naming schemes. I know
this will change over time, but I'm thinking of users in
the near term who will want to have access to OTS datasets
from different sources.

Since the OTS data management team started by deciding to
use NetCDF, CF was a natural choice; the CF standard
includes metadata features and syntax that we wanted to
incorporate, besides the standard names. I am not sure
what we would gain by adopting CF without using their
vocabulary. That vocabulary is specifically tailored
to NetCDF; the nature of NetCDF allows us to have a
reasonably short list of standard names, modified by
attributes such as units and long_names, and allowing
additional attributes to describe things like data source
and processing methods.

The BODC vocabulary, which I understand was developed
for use with a database, necessarily limits the use of
modifying attributes, and combines much of this information
into a single field; hence the extremely long list of
parameters. The BODC scheme is excellent for data base
use, but it may not take full advantage of the flexibility
that NetCDF attributes offer. This may make mapping more
of a challenge for software, but it will certainly make
it difficult for the average user or data provider to do
by looking at the two lists.

> The weakness with this is that the structure implies
> the relationship:
>
> BODC Parameter Usage Vocabulary term = (ie is exactly
> the same as) CF Standard Name term

I agree with Roy that this is a big problem. The details
embedded in the BODC parameter codes would make it difficult
for a user to select a code from his table that is really
accurate for an OTS parameter. I have spent some time
trying to compare the BODC codes with the CF standard_names
for our parameters, as I'm sure others have done. I wasn't
able to do this very successfully, maybe because I wasn't
looking at the best presentation of the BODC dictionary.

Is there a web page with a formatted presentation of the
BODC codes, for those of us who don't use Microsoft Access?
I'd like to try to see how difficult it is to pick accurate
BODC codes for the variables we measure.

Thanks -
Nan


*************************************************************
*Nan Galbraith Upper Ocean Processes Group *
*Woods Hole Oceanographic Institution Woods Hole, MA 02540 *
*************************************************************
Received on Fri Jun 17 2005 - 09:47:47 BST