[CF-metadata] file with both run time and forecast (valid) time coordinates

From: Jonathan Gregory <j.m.gregory>
Date: Wed, 6 Sep 2006 07:27:07 +0100

Dear John

> One question on situation (b): You make both one-dimensional auxiliary
> coordinate variables, but could you make one a coordinate and one an
> auxiliary coordinate? Eg:

Coordinate variables have to be monotonic, and this will not generally be
true in the cases where you have, in effect, a 2D (possibly sparse) set of
combinations of analysis and validity times. That is the reason for using
an index axis, with both times as auxiliaries.

Best wishes

Jonathan
Received on Wed Sep 06 2006 - 00:27:07 BST