Hi Eizi,
Thank you for this clarification.
(I was aware of the UKMO proposal to expand GRIB2 for UK National Grid -
we are in the process of formalising our internal governance of data
formats, and so the proposals for the WMO InterProgramme Expert Team on
Data Representations & Codes (IPET-DRC) - both our own and others -
received wider review than has previously been the case.)
Regards,
Bruce
-----Original Message-----
From: TOYODA Eizi [mailto:toyoda at gfd-dennou.org]
Sent: 06 October 2011 13:30
To: Wright, Bruce; Bryan Lawrence; cf-metadata at cgd.ucar.edu
Subject: Re: [CF-metadata] Question on WKT representation of CRS
(Bentley,Philip)
Hello Bruce,
Yes there is much overlap between CRS informaton and existing CF
attributes for grid mapping.
The only missing thing is identification of datum. Shape of the
sphere/spheroid is not sufficient, since the same spheroid is often used
in many datums with location shift. For example, EPSG:6277 (OSGB 1936
datum) in WKT is following:
DATUM["OSGB_1936",
SPHEROID["Airy 1830",6377563.396,299.3249646,
AUTHORITY["EPSG","7001"]],
AUTHORITY["EPSG","6277"]]
All numeric parameters can be expressed by inverse_flattening and
semi_major_axis, but the two attributes indicates only spheroid (Airy
1830).
The name "OSGB_1936" is essential to distinguish the datum from (for
example) Irish ones.
You might know UKMO proposed to expand GRIB2 for UK National Grid
recently.
I commented the same point, and the WMO discussion converged to add
datum name. So similar approach is suggested if we take conservative
approach to expand existing grid mapping attributes.
Best,
Eizi
----- Original Message -----
From: "Wright, Bruce" <bruce.wright at metoffice.gov.uk>
To: "Bryan Lawrence" <bryan.lawrence at ncas.ac.uk>;
<cf-metadata at cgd.ucar.edu>
Sent: Thursday, October 06, 2011 1:42 AM
Subject: Re: [CF-metadata] Question on WKT representation of CRS
(Bentley,Philip)
> Hi Bryan,
>
> I suspect it probably is becoming important for weather forecast
> output now.
>
> At the Met Office, we're now generating 2km post-processed forecast
> data (from models running at up 1.5km resolution). At present, this is
> mainly shared internally (and with a small number of 'expert'
> customers) using bespoke data formats (which don't fully describe the
> CRS) . However, in future I'd expect to be sharing these data more
> widely, ideally using CF-netCDF (and GRIB2), and would feel more
> comfortable, it we could fully characterise CRS in use, to ensure
> appropriate use in combination with other data.
>
> But to echo your last comment, a CRS WKT enhancement should be
> optional, and not replace the current CF grid description information
> - we don't want to break existing client software, just allow a fuller
> description of the CRS where required.
>
> Regards,
> Bruce
> --
> Bruce Wright Expert IT Analyst (Data Management) Met Office FitzRoy
> Road Exeter EX1 3PB United Kingdom
> Tel: +44 (0)1392 886481 Fax: +44 (0)1392 885681
> E-mail: bruce.wright at metoffice.gov.uk http://www.metoffice.gov.uk
>
> -----Original Message-----
> From: cf-metadata-bounces at cgd.ucar.edu
> [mailto:cf-metadata-bounces at cgd.ucar.edu] On Behalf Of Bryan Lawrence
> Sent: 05 October 2011 12:52
> To: cf-metadata at cgd.ucar.edu
> Subject: Re: [CF-metadata] Question on WKT representation of CRS
> (Bentley,Philip)
>
> Hi Seth
>
> I haven't read all the threads ... but I strongly agree with your
> last paragraph!
>
> I have had many conversations with folks who think that adding datums
> will make data more usable to the impacts community, where datum
> errors can move things by o(10)s of km ... and my protestation that no
> one should interpret as physical any differences on those scales from
> a
> (climate) model (even one run at o(km) resolution if such exists) ...
> was simply ignored. The reality is exactly what you say, that level of
> specificity is simply inappropriate.
>
> I appreciate some of the arguments raised in the thread on storing
> lat/lon coordinates, about the need for the use of one in a GIS
> workflow
> - but frankly I think that's an issue about workflow metadata not
> source data metadata. As Balaji and others said, there might not even
> be *one* datum appropriate for GCM work ...
>
> Of course observational data may well be different, and I'm not sure
> about NWP ... especially mesoscale models. So by all means, facilitate
> the provision of this information, but don't make it compulsory ...
> and I think it would be with WKT?
>
> Cheers
> Bryan
>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
Received on Thu Oct 06 2011 - 09:55:32 BST
