[CF-metadata] Example of forecast data

From: Brian Eaton <eaton>
Date: Wed, 16 Jul 2003 11:20:27 -0600

Hi Jonathan and Ag,

I'm opposed to adding forecast_validity_time to the standard name table for
the following reasons:

The purpose of standard names is to identify quantities and to help users
of data from different sources decide whether or not certain variables
contain data that is comparable. The name forecast_validity_time does not
help in this regard. forecast_validity_time means exactly the same thing
as "time", so adding that name makes it possible to say the same thing
using more than one standard name. This makes things more complicated
without adding information.

That data may have been produced by a forecast is a property of the data
and not of the time coordinate. Same is true of an analysis or
observational data. The time at which a forecast, analysis, or observation
is valid should be indicated by an axis with the standard name "time".

The standard name table currently contains only one other name which refers
to an instant in time, i.e., "forecast_reference_time". It is sufficient
to have these two names to distinquish between valid and reference times of
a data variable.

Brian
Received on Wed Jul 16 2003 - 11:20:27 BST