[CF-metadata] water level with/without datum

From: Jonathan Gregory <j.m.gregory>
Date: Fri, 26 Feb 2010 09:41:55 +0000

Dear Nan

> >I think Roy's example is a relevant use case. Although he has not made a
> >proposal, his data set requires either a new name of
> >river_water_temperature,
> >or a name which can be used for both sea and river. The existing name of
> >sea_water_temperature is not sufficient for the case he described.
>
> Roy's example shows the need for a *single name* that can be used for
> both sea and river temperature, not different names, if I understand his
> description correctly.

That's right. He needs a single name which covers sea and river.

> I don't see any reason to (or reasonable way to) split up the measurements
> made by these instruments based on which side of an invisible line they're
> on at any given point.

Yes. In some applications it is artificial to distinguish between sea, lake,
reservoir, river etc.

> There are several names that use the modifiers 'atmosphere', 'in_air' and
> 'surface' to indicate water that's not part of a water body. Does this imply
> that the unmodified term 'water' means water that's in a water body?
>
> The only names I can find that use plain 'water' seem to be sound_intensity
> and sound_pressure terms - I assume these refer to water in a water body?
> Is that enough of a precedent to suggest that water_temperature, _velocity,
> _salinity, etc etc could be standard names for properties of the water in
> bodies of water?

I don't remember the intention of those standard names. I am surprised that
they don't say "sea" or "river" if that's what they meant.

I am sorry to be obstinate on this, but I don't think that it would be right
to assume that "water" without any qualifier meant water that is part of a
water body. As I said in a previous email, we always try to indicate the
context explicitly in standard names, to make them self-describing. This use
of "water" would be a kind of definition by omission, rather than explicitly.

I think that our discussion so far indicates that we should keep the existing
"sea" names, and add corresponding new names for water bodies in general (as
and when they are requested). We can likewise add new "sea" names.

Therefore I think we have to decide what to call the new names. Roy suggested
water body. As I've said before, I would prefer sea/lake/river_water (or with
some other punctuation) to water_body_water, because sea/lake/river_water is
more self-explanatory, and the repetition of "water" in water_body_water is
clumsy and possibly confusing. I can imagine someone not being sure how to
parse "water body water temperature" when they first come across it.

As Roy said, "/" could be a problem, although it's legal for netCDF. It does
make the intention clearer, since "/" means "or". It could be spelled out,
at the cost of greater length, as sea_or_lake_or_river_water.

best wishes

Jonathan
Received on Fri Feb 26 2010 - 02:41:55 GMT