[CF-metadata] fixed sensors, depth, datum

From: Jonathan Gregory <j.m.gregory>
Date: Mon, 15 Sep 2008 18:32:28 +0100

Dear all

Referring to Dale's list of requirements

- height relative to the ellipsoid
- height relative to the geoid
In my opinion these are distinct geophysical quantities. As Dale pointed out,
we already have the standard name of "altitude" for the second of them, and
the first should probably have the name height_above_reference_ellipsoid for
consistency with other standard names. That would need to be proposed to this
email list.

- name of the ellipsoid
- name of the geoid
These could be added as possible attributes to the grid_mapping variable.
That is a conventions change that would need a trac ticket to propose and
discuss. In fact Phil Bentley did propose to have such attributes in his
CRS ticket. We had discussions in which I expressed my usual disquiet about
redundancy. What if the name of the ellipsoid is not consistent with the
parameters of the ellipsoid that have been supplied, if they have been? I
am not sure how to resolve that. This concern does not apply to the geoid,
because it is not practical to specify it in detail as part of CF metadata;
we can only identify it by name. If the ellipsoid_name is a grid_mapping
attribute too, it simplifies Dale's example, as it doesn't have to appear
in its own dummy variable.

- altitude of (offset to) other common datums (e.g. MLLW)
Dale's offsets are single numbers, but they could in some circumstances be
(lon,lat) fields, as Ethan says. In either case, they can be data variables,
so I see no problem that that. The other common datums are tidal datums, for
instance. Again, I think the various tidal levels are distinct geophysical
variables, and need different standard names to be proposed
e.g. altitude_of_mean_lower_low_water and altitude_of_mean_low_water.

I think I don't properly understand how these concepts relate to concepts
like "Australian height datum", which Ethan mentions. What does that mean
exactly, can someone explain? I suspect it must mean the designation of a
particular geoid, and saying that a particular location has a certain altitude
wrt that geoid. By specifying a geoid, a location and an altitude you would
thereby define altitude for everywhere else in the world. Is that right?

Best wishes

Jonathan
Received on Mon Sep 15 2008 - 11:32:28 BST