Hi.
The new grid_mapping scheme looks great! It accomplishes, in a more backward compatible way, most of the things that I was suggesting.
Regarding Jonathan's reply to my previous comment:
If I read a temperature using a thermometer, the value does not have a geographic coordinate system associated with it. If I also use my handy GPS unit to read my location, I can associate that location with the temperature. The location does have a coordinate system associated with it, and in fact is nearly meaningless without one. I could also use more traditional surveying tools and determine a location with respect to an entirely different coordinate system. The coordinate systems naturally associate with the locations, not the temperature. The current CF approach leaves the coordinate system association with a given regular or auxiliary coordinate variable unspecified (in any direct fashion). In order to determine the coordinate system for a coordinate, you must search the data variables for an association with the coordinate (through shape or through the coordinates attribute), then find the grid_mapping attribute. The new scheme represents an improvement, but I still find myself asking why we
are placing a fundamental attribute of a coordinate anywhere other than directly with the coordinate.
Grace and peace,
Jim
Jim Biard
Research Scholar
Cooperative Institute for Climate and Satellites
Remote Sensing and Applications Division
National Climatic Data Center
151 Patton Ave, Asheville, NC 28801-5001
jim.biard at noaa.gov
828-271-4900
On Sep 5, 2012, at 1:28 PM, Jonathan Gregory <j.m.gregory at reading.ac.uk> wrote:
> Dear Randy
>
> I agree with what you write. You could list the alternative sets of three
> coordinates each all as auxiliary coord vars, and distinguish them with
> standard_names; presumably you would want to propose some new standard_names.
> That would be fine and quite simple to do.
>
> In addition, you could propose new grid_mappings. Up to now grid_mappings
> have been for 2D coordinate systems, but you would want something in 3D,
> I think. If it's likely generally useful, it would interesting to see a
> proposal of what it might look like.
>
> There is already an agreed change to the CF standard to allow more than
> one grid_mapping for a given data variable. This was proposed by Mark Hedley
> in ticket 70 https://cf-pcmdi.llnl.gov/trac/ticket/70 and will go in CF 1.7.
>
> In response to Jim's comment, I would say that the data variable does have a
> coordinate system; it's defined by the various coordinate variables that the
> data variable is associated with, either 1D (Unidata) coordinate variables
> or auxiliary coordinate variables. The role of the grid_mapping is to make
> explicit the relationship among the coordinate variables of the space in which
> the data variable exists.
>
> Best wishes
>
> Jonathan
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <
http://mailman.cgd.ucar.edu/pipermail/cf-metadata/attachments/20120905/799dde4b/attachment.html>
Received on Wed Sep 05 2012 - 12:29:00 BST
