Hi Andreas,
In CF the rule about the units attribute is that it has to be a string that is recognized by the UDUNITS package -
http://www.unidata.ucar.edu/software/udunits/. Provided UDUNITS knows how to convert between the unit used in the file and the canonical units specified in the standard name table then the CF conventions are satisfied. The issue with using molecules cm-2 is that UDUNITS does not currently contain an entry for a unit of molecules although it does contain an entry for mole. (The list of currently supported units is held in a series of XML files:
http://www.unidata.ucar.edu/software/udunits/udunits-2-units.html.)
If we can get molecule added as a unit to UDUNITS along the lines suggested by Sander earlier in the discussion, then I think you could use molecules cm-2 in your files and we could still have mol m-2 as the canonical unit.
Sander Niemeijer wrote:
>
> <unit>
> <def>mol/6.02214179e23</def>
> <name>
> <singular>molecule</singular>
> <plural>molecules</plural>
> </name>
> <aliases>
> <name><singular>molec</singular></name>
> </aliases>
> </unit>
If you are happy for us to proceed down this route then I think the names could be accepted now and added to the standard name table with canonical units of mol m-2. I will then send a request to ask if molecules can be added to UDUNITS (although I'm not sure of the time scale for adding the new unit). Would that be an acceptable way forward?
Best wishes,
Alison
------
Alison Pamment Tel: +44 1235 778065
NCAS/British Atmospheric Data Centre Email: alison.pamment at stfc.ac.uk
STFC Rutherford Appleton Laboratory
R25, 2.22
Harwell Oxford, Didcot, OX11 0QX, U.K.
> -----Original Message-----
> From: Andreas Hilboll [mailto:lists at hilboll.de]
> Sent: 18 March 2013 11:15
> To: Pamment, Alison (STFC,RAL,RALSP); cf-metadata at cgd.ucar.edu
> Subject: Re: [CF-metadata] how to specify time dimension for monthly
> averages
>
> > I'm not sure if you noticed my post to the CF mailing list last week about
> the tropospheric/stratospheric trace gas standard names you proposed back
> in December: http://mailman.cgd.ucar.edu/pipermail/cf-
> metadata/2013/056207.html. There is just one question remaining to be
> resolved about the units - whether you are happy with the consensus view
> that we should use mol m-2? I am hoping to make an update to the standard
> name table later this week and it would be great if we are in a position to
> include your names. Please could you let me know via the mailing list how
> you would like to proceed?
>
> Hi Alison,
>
> thanks for the reminder; somehow this mail didn't get through to me.
>
> In atmospheric trace gas remote sensing, nobody to my knowledge uses the
> unit "mol m-2".
>
> General question: Is it "allowed" to use a different unit in the
> variables metadata than is specified in the standard name table? I.e. if
> the new standard name table contains "mol m-2" as unit, would it still
> be "legal" for me to use "molecules cm-2" in the file itself?
>
> Best regards, Andreas.
--
Scanned by iCritical.
Received on Mon Mar 18 2013 - 05:56:21 GMT
