[CF-metadata] Surface Air Temperature

From: Roy Lowry <rkl>
Date: Fri, 10 Oct 2003 07:33:24 +0100

Dear Jonathan,

In your reply to Curt you state

'The standard name of "air_temperature" should be used, with a scalar
coordinate
variable or a size-one dimension of height, specifying 1.5 m, 2 m or
whatever
it happens to be.'

Why not take the same approach to sea surface temperature and sea
surface salinity? z=0 could be used for skin temperatures (ie
satellite
data or its modelled equivalent) and z= whatever (I guess a negative
number as height is a more natural concept than depth) for
bucket/thermosalinograph/near-surface model T/S.

Differentiation between surface and subsurface T/S goes back to the
work on GF3 in the 1980s and causes headaches when trying to generate
3-D observational data syntheses and usually results in
thermosalinograph data, much of it good stuff, being excluded. I've
unified the coding in BODC, although surface data my usually be
inferred
from method codes as different codes are used for thermosalinograph
and
CTD data (hence a mapping to differentiated codes is possible).

As a dictionary guardian I fully appreciate the problems in changing
standard name lists, particulary when it means the loss of a term that
is in use in the community. However, would an evolutionary process be
acceptable with the CF community encouraged to treat T/S the same as
air
temperature whilst not eliminating the surface terms from the standard
list?

Cheers, Roy.
Received on Fri Oct 10 2003 - 00:33:24 BST