[CF-metadata] CF feature type trajectory (Ch. 9; May 10, 2011) and axis, attribute

From: Jonathan Gregory <j.m.gregory>
Date: Tue, 26 Jul 2011 10:13:51 +0100

Dear Chris, Jeff, Steve, John

> > > It seems to me that z in the examples of Appendix H (formerly
> > > designated A9) plays the role of an auxiliary coordinate variable,
> > > although technically it's not because there isn't (but should be) a
> > > 'coordinates' attribute which lists auxiliary coordinate variables.
> > > I don't know why axes are banned for auxiliary coordinate variables in
> > > the case of trajectories. Something needs to be clarified here, or
> > > the examples changed.
> > >
> > > This was ticket 37. Are there any comments from Steve Hankin, the
> > > moderator, or any of the other people who contributed to writing
> > > Chapter 9, or anyone else who knows more than me?
> stand by. I think Chris has pointed out a needed correction to Chapter
> 5 that was discussed, but was not documented. (thanks Chris) But need
> to confirm with others.

Ticket 62 discusses the proposed change to allow auxiliary coordinate
variables to have the axis attribute. This ticket was originally proposed by
Karl to correct a defect, and I think when we were writing chapter 9 we had
in mind this change. I propose that ticket 62 should be implemented in CF 1.6
(it wasn't one of the tickets I had listed earlier to Jeff - I must have
overlooked it). In that ticket, I proposed some wording, to which no-one
subsequently objected, and that was many months ago.

Also, Jeff, if there are any examples in App H where a data variables does
not list its aux coord vars in a coordinates attribute, please could you
correct them? This is required, as you say. It says in 9.5, "The coordinates
attribute must be attached to every data variable to indicate the
spatiotemporal coordinate variables that are needed to geo-locate the data."

Best wishes and thanks

Jonathan
Received on Tue Jul 26 2011 - 03:13:51 BST