All,
IMHO John's question is part of a broader issue in the construction of standard
parameter names -- the need for hierarchy. (I confess at the outset that these are
inchoate thoughts ... so this email is just to open the dialog.)
Looking at names like
tendency_of_air_temperature_due_to_diabatic_processes
tendency_of_air_temperature_due_to_dry_convection
tendency_of_air_temperature_due_to_large_scale_precipitation
...
we see how a single conceptual quantity, "tendency_of_air_temperature", has diverged
into many variables. Some important semantics have been lost in the process -- that
these variables are part of a family (terms in a "budget"). Simply knowing that the
variables share the same units is not a sufficient indication (to a machine) that it
is reasonable to allow these variables to be added together.
John's example is related. A "delta" of a quantity is still that same quantity, but
with some (possibly time/space dependent) offset removed. This kind of information
needs to be captured in a sophisticated naming scheme (but how best to do so?).
This email is more trouble than help, I know. But that seems to be the nature of
the beast.
- steve
==================================
Brian Eaton wrote:
> I agree with Jonathan that the distinction should be made in the
> description of the quantity, and not in the units. We do have the standard
> name air_temperature_anomaly, but anomaly implies that the difference is
> with a climatology. If we had a standard name like
> air_temperature_difference it seems that would be appropriate. But that
> leaves an unresolved issue as to what the difference is with respect to.
> Of course the same issue is present with an anomaly; we don't currently
> have a mechanism for connecting an anomaly with the appropriate
> climatology.
>
> In view of the problems with interpreting exactly what a difference field
> means, the question that comes to mind is why is WRF writing temperature
> differences to the output file? Why not convert to a standard quantity
> before output (or during conversion to CF compliant output)?
>
> Brian
>
> On Tue, Aug 10, 2004 at 10:49:33PM +0100, Jonathan Gregory wrote:
> > Dear John
> >
> > > we're looking at WRF model output with temperature differences in
> > > Kelvin. unfortunately, our software package converts to Celsius by
> > > subtracting 273. We could fix if we knew it was a temp difference
> > > instead of an absolute temp.
> >
> > I don't think the units would be able to help you. Does the WRF model output
> > distinguish in any other way between temperatures and temperature differences?
> > Would your software be able to look at other attributes to decide what to do?
> >
> > Cheers
> >
> > Jonathan
> > _______________________________________________
> > CF-metadata mailing list
> > CF-metadata at cgd.ucar.edu
> > http://www.cgd.ucar.edu/mailman/listinfo/cf-metadata
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://www.cgd.ucar.edu/mailman/listinfo/cf-metadata
--
???,??,????`????,??,????`????,??,????`??????,??,???
Steve Hankin, NOAA/PMEL -- Steven.C.Hankin at noaa.gov
7600 Sand Point Way NE, Seattle, WA 98115-0070
ph. (206) 526-6080, FAX (206) 526-6744
Received on Mon Aug 16 2004 - 18:31:11 BST
