[CF-metadata] geoid

From: Jonathan Gregory <j.m.gregory>
Date: Mon, 17 Nov 2003 09:08:44 +0000

Dear Brian

> If we take the path of including information that identifies the geoid in
> standard names then the problem arises of providing a new standard name for
> each geoid or horizontal datum reference of interest. It seems better to
> keep this information out of the standard name and provide separate
> attributes for it since it's really part of defining the spatial coordinate
> system.

I agree.

> For defining geodetic models the FGDC doc
> provides horizontal_datum_name and ellipsoid_name to identify published
> references, e.g., "North American Datum of 1983", and "Geodetic Reference
> System 80". It seems that we should allow these attributes to be attached
> either to variables containing data which requires knowlege of the geoid
> for correct interpretation, or to variables that define map projections.

Is this a proposal for new CF attributes of these names? Can we point to FGDC
for their values or should we incorporate them in a new CF list? I suppose
it depends whether FGDC is guaranteed to be comprehensive.

Such attributes could be required for
* grid_mapping variables.
* data variables of quantities that depend on geoid, such as sea surface
elevation.
* vertical coordinate variables defined wrt geoid, such as altitude.

Presumably we leave it open to the data-writer to decide when the geoid needs
to be stated.

Cheers

Jonathan
Received on Mon Nov 17 2003 - 02:08:44 GMT