Hello Mark,
Thanks for presenting this clearly. I favour the second option of an
expanded syntax for grid_mapping attribute on a data variable.
Pragmatically, I think that it is possible for two data variables in
the same file to share identical latitude and longitude coordinate
values but have different grid_mappings. This would occur, for
example, if one data variable came from a climate model which used the
spherical earth approximation and the other used a spheroidal
earth. In this case, if the coordinate reference information was
attached to the coordinates I would need to duplicate the coordinates'
data in my file, which is not desirable.
> * a view that the coordinate is defined with respect to a coordinate
> system and a data variable is defined with respect to a coordinate
Conceptually, I prefer a different interpretation of the role of
coordinate systems. I like to think of a coordinate reference system
as being defined by the arbitrary coordinates which describe the
dimensions of a space (or grid) in which a data variable exists. From
this view point, hanging space-defining attributes off the data
variable seems more natural to me.
All the best,
David
--
David Hassell
National Centre for Atmospheric Science (NCAS)
Department of Meteorology, University of Reading,
Earley Gate, PO Box 243,
Reading RG6 6BB, U.K.
Tel : 0118 3785613
Fax : 0118 3788316
E-mail: d.c.hassell at reading.ac.uk
Received on Thu Nov 03 2011 - 04:05:34 GMT
