[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[THREDDS #FUB-114946]: Understanding vertCoord behaviour with netcdf subset service

Hi Daryl,

I started to take a look at the phenomenon you are seeing, but I immediately 
noticed that the Thredds server you've referenced is running a very old version 
of the TDS - 4.2.10 from April of 2012.  We've done a lot of work since then, 
and there is a reasonable good chance that a bug contributing to the behavior 
has been fixed in a subsequent version of the TDS.

Any chance you can convince them to upgrade?  We're gone all the way through 
4.3.20, and are about to release 4.4.0.

Cheers,
  Lansing Madry
  Unidata
  Boulder, Colorado

> Hi Thredds Dudes,
> 
> I was recently exchanging emails with NCDC about their thredds server and
> how it was handling vertCoord requests for the netcdf subset service.  I
> am really confused by the behaviour and perhaps there are bugs afoot here!
> Lets start this goose chase with the documentation:
> 
> http://www.unidata.ucar.edu/software/thredds/current/tds/interfaceSpec/NetcdfSubsetService_4_3.html#Vertical
> 
> "Subset on the vertical axis of a variable or variables with the same
> vertical levels may be done with the vertCoord parameter"
> 
> Okay, so if I go here:
> 
> http://mtarchive.geol.iastate.edu/thredds/ncss/grid/mtarchive/2013/11/04/gempak/model/2013110412_avn211.gem/pointDataset.html
> 
> and put in '2' for the vertical level and TMPK_HGHT var, I get
> 
> http://mtarchive.geol.iastate.edu/thredds/ncss/grid/mtarchive/2013/11/04/gempak/model/2013110412_avn211.gem?var=TMPK_HGHT&latitude=41.99&longitude=-93.61&temporal=all&time_start=2013-11-07T18%3A00%3A00Z&time_end=2013-11-14T12%3A00%3A00Z&time=2013-11-07T18%3A00%3A00Z&vertCoord=2&accept=csv&point=true
> 
> lots of NaN !  Hmmm, if I change, vertCoord to blank
> 
> http://mtarchive.geol.iastate.edu/thredds/ncss/grid/mtarchive/2013/11/04/gempak/model/2013110412_avn211.gem?var=TMPK_HGHT&latitude=41.99&longitude=-93.61&temporal=all&time_start=2013-11-07T18%3A00%3A00Z&time_end=2013-11-14T12%3A00%3A00Z&time=2013-11-07T18%3A00%3A00Z&vertCoord=&accept=csv&point=true
> 
> still NaN...  If I then change vertCoord to 0, I get data!
> 
> http://mtarchive.geol.iastate.edu/thredds/ncss/grid/mtarchive/2013/11/04/gempak/model/2013110412_avn211.gem?var=TMPK_HGHT&latitude=41.99&longitude=-93.61&temporal=all&time_start=2013-11-07T18%3A00%3A00Z&time_end=2013-11-14T12%3A00%3A00Z&time=2013-11-07T18%3A00%3A00Z&vertCoord=0&accept=csv&point=true
> 
> but then if I add UREL_HGHT
> 
> http://mtarchive.geol.iastate.edu/thredds/ncss/grid/mtarchive/2013/11/04/gempak/model/2013110412_avn211.gem?var=TMPK_HGHT&var=UREL_HGHT&latitude=41.99&longitude=-93.61&temporal=all&time_start=2013-11-07T18%3A00%3A00Z&time_end=2013-11-14T12%3A00%3A00Z&time=2013-11-07T18%3A00%3A00Z&vertCoord=0&accept=csv&point=true
> 
> I get data, but the vertCoord[unit="m"] is now zero for everything.
> 
> Looking back at the documentation, it has:
> 
> Request: "Give me variable Temperature_pressure at a particular vertical
> level: 1000 mb*"
> Temperature_pressure                                          1000
> ?var=Temperature_pressure&vertCoord=1000
> 
> so it would appear vertCoord is the value along the vertical axis and not
> the index along the axis?
> 
> Any thoughts on this?  I understand that this is very complex problem to
> account for.
> 
> thanks,
> daryl
> 
> --
> /**
> * Daryl Herzmann
> * Assistant Scientist -- Iowa Environmental Mesonet
> * http://mesonet.agron.iastate.edu
> */
> 
> 


Ticket Details
===================
Ticket ID: FUB-114946
Department: Support THREDDS
Priority: Normal
Status: Open