[CF-metadata] units for sea water salinity

From: Jonathan Gregory <j.m.gregory>
Date: Wed, 4 Aug 2010 09:55:24 +0100

Dear Upendra

> This is in regard to the units corresponding to the standard name
> sea_water_salinity. We here at NODC use PSS, which is unitless, for
> Salinity. PSU is not a good way to represent salinity and shouldn't be
> used. The units for sea_water_salinity should be changed to PSS in the
> standard name table.
>
> On a related note, should I be using Standard Name at all if my units
> are different from those in the table?

There is more than one point here.

(1) PSU (and PSS) are not udunits, and for that reason the canonical unit
given in the standard name table is actually 1e-3 (dimensionless). For a long
while we have had an intention to produce a CF variant of the udunits table
to include PSU, but not done it yet.

(2) In climate model data, which is an important (and the original)
application of CF, the exact definition of salinity is not so relevant,
because there are lots of approximations in the models, as you know.

(3) However, CF standard names in many cases provide different degrees of
precision for different purposes. Distinctions which might not be made in one
application are crucial in another. Perhaps you might propose some
alternative standard_names to sea_water_salinity, with more precise
definitions, that would meet your needs and could be included in the table
as well as the existing more imprecise name, which is fine for climate
modelling and other purposes. The strict answer to your question is No,
if your unit is not dimensionally consistent with the canonical unit in the
standard table, you should not use that standard name.

Best wishes

Jonathan
Received on Wed Aug 04 2010 - 02:55:24 BST