Hi -
And just to clarify one thing about Udunits. Ferret does not link with
the udunits library but is consistent
with the udunits standard for the set of units that are outlined in the
COARDS and CF
Standards. It does not have internal conversions for all of the units
defined in the udunits standard. So it will store the units string
units
and will allow you to make use of it in your scripts, but in this
instance it doesn't
convert Pascales to other units of pressure.
So you'll see messages from Ferret telling you that the units are not
recognized and will not be convertible. This means that there is no
internal logic to do conversions, as there is for calendar time units,
or degrees-longitude which may be given as degrees East or degrees
West, which Ferret can reconcile with other calendar or longitude axes,
or as Steve says for handling integrals and derivatives on
longitude/latitude axes.
The warnings say,
yes? define axis/units=pa/z/depth zax = {1,2,3}
*** NOTE: (fyi) units are not subject to auto-conversion: PA
yes? use m_data.nc
*** NOTE: Units on axis "ZAX" are not recognized: PA
*** NOTE: They will not be convertible:
Ansley
Steve Hankin wrote:
Hi
Patrick,
Thanks, as ever, for your always-useful feedback. Some discussion of
this point ...
Ferret should never "reject" units on axes. (If you have a case where
it does, lets follow up on that example.) It should merely inform you
that it does not how to perform conversions automatically on those
units. Unit conversions are applied by Ferret to standardize the
computation of derivatives and integrals along axes. Since the
conversion of Z axis units between pressure units and depth/elevation
units is data-dependent, Ferret cannot perform it in a general way.
One could imagine 2 tracks of convertibility for Z axis units -- depths
convertible between (say) meters and centimeters; pressures similarly
convertible among only pressure units. Are there use cases to motivate
the effort of this change? (Or is the best solution merely to tone
down the strength of the warning about non-convertible units?)
- Steve
Brockmann Patrick wrote:
Hi all,
Why Pa unit is not accepted as a pressure unit by ferret
whereas it is well present in the file
http://www.unidata.ucar.edu/software/udunits/udunits-1/udunits.txt
Patrick
|