[CF-metadata] mixing ratio

From: Jonathan Gregory <j.m.gregory>
Date: Fri, 7 Nov 2008 08:39:05 +0000

Dear Martin

Your arguments are reasonable and we've had this kind of discussion a few
times about what standard names are for. I agree with what you write:

> I suggest that standard names should
> be a reasonable compromise between accuracy and simplicity. We should
> definitively avoid jargon and make an effort to come up with defintiions
> that are as precise and clear as possible, but it should be recognized
> that some prior knowledge might be required before the full meaning of a
> term can be understood.

Of course, it depends on where the compromise should be struck. I would point
out also that

* Standard names are not really "names". We do not necessarily use the terms
which are shortest and most familiar, because they may be jargon. Standard
names are, I think, a very brief answer to the question, "What does that
mean?". No-one is suggesting that people would use standard names in
conversation or in scientific literature. That is not their purpose.

* The target audience is any user of the data. All users have some general
physical expertise, but not necessarily in the specific field. The names
should at least indicate to any user which general area they refer to.

* The names have to be able to make sufficient distinctions so that users of
different datasets can decide whether the quantities in them are comparable.
That means they may have to spell out more than you would in the context of
a particular model. That's the reason for spelling out "epineutral", "tracer"
and "biharmonic" in the example you quote. These words could not be replaced
with self-explanatory terms unless we had a far longer name, so we depend on
the definition or specific expertise for that. But at least the name indicates
to you that it is an ocean quantity and a diffusivity, and that's useful.

Back to the specific point. If the distinction between ambient air and dry
air is actually of no importance for the cases we are currently dealing with,
I agree that mass_fraction_of_X_in_air for all X is the best choice, and
the definition can say that it is not specified whether "air" is dry or
ambient. That would be fine.

If a subsequent use-case arises where someone does need to make the distinction
we will introduce more precise names which indicate that. That would not be
a replacement, but an addition to the standard name table. The use of standard
names depends on the application; some have to be more precise than others.

Best wishes

Jonathan
Received on Fri Nov 07 2008 - 01:39:05 GMT