Hi all -
On 9/8/11 11:54 AM, Jonathan Gregory wrote:
> ... The raw and corrected variables have the same standard name
> because they are the same geophysical quantity, but you can
> distinguish them for your own purposes with other non-standardised
> atts such as long_name. So far there has not been a proposal to
> standardise this kind of distinction.
I agree that this approach is legal in CF, but it doesn't help make
data interoperable or machine-discoverable. Another way to
approach this, especially if your file is intended to be shared,
is for the one variable that best represents the concept of easting
to have the CF standard name projection_x_coordinate. Any
other variables that are "alternate" versions of that can be
included without a standard name - this is also completely legal
in CF.
In OceanSITES, we want to try to make only the "best" version
of any data discoverable, while still providing access to raw
versions or to versions that are required for provenance. Our
workaround for this is to remove the standard name from all
but the single "primary source" - the authoritative record of
a particular variable.
IMHO, this approach will work better in the long run.
- Nan
> ... I map 'Easting" to CF convention:'projection_x_coordinate', where
> easting represents raw UTC coordinates.
>
> However, I also have 'CorrEasting' which represents the corrested
> (despiked, smoothed) easting for the same observation.
>
--
*******************************************************
* Nan Galbraith (508) 289-2444 *
* Upper Ocean Processes Group Mail Stop 29 *
* Woods Hole Oceanographic Institution *
* Woods Hole, MA 02543 *
*******************************************************
Received on Fri Sep 09 2011 - 06:17:59 BST
