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

20040226: IDV - WIN2000 - netCDF files from ECMWF

>From: "Dr. Hans-Juergen Panitz" <address@hidden>
>Organization: Forschungszentrum Karlsruhe / IMK, Germany
>Keywords: 200402260822.i1Q8MCF9017928  IDV MARS convention

Hi again-

>Institution:  Forschungszentrum Karlsruhe / IMK, Germany
>Package Version: 1.1b2
>Operating System: WIN2000
>Hardware Information: 1 GB Processor
>Inquiry: I downloaded a 3-D  netCDF file (1day, 6-hourly air temperatures on p
> ressure levels) from the ECMWF ERA-40 server (40 years ECMWF reanalyses).
>
>The IDV cannot work with the data.
>After  choosing the file the button "3D Grid"
>shows the entries "temperature" and  "Derived".
>Taking "Temperature" the pull down menue to choose the different chart types o
> pens.
>
>Choosing anyone of these alwas the message
>
>"Unable to make a Field from data with parameter name :t"
>
>What is the problem? 
>Can't IDV be applied with ECMWF dats?

As I mentioned in my previous note, all netCDF files are not created
equal.  In the case of this file, it uses the "MARS" convention which
we currently have no support for.  The problem with this file is
the time specification.  If you look at the time variable, it is
listed as:

        int time(time) ;

with no units.  The values are:

time = 0, 600, 1200, 1800 ;

which are really in units of centihours from some reference point.
The reference point is given by the global attribute "date":

                :date = "20000801" ;

(presumably Aug 1, 2000).

In other conventions (e.g., CF) the time coordinate variable is
usually self contained (e.g.,  seconds since 1970-01-01 00:00)
and we can parse that. 

We could probably add a handler for the MARS convention, but I could
not find any documentation on the ECMWF website about this convention.
Do you have access to such documentation?

In the mean time, I am investigating another possibility using
the netCDF Markup Language (NcML).  Your file is a good test 
case for the IDV to test it's use.  I will let you know if I
have success.

>I attached the ncdump output of my test file.

Thanks, that was useful.

Don Murray
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
303 497 8643                                                  P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata WWW Service              http://my.unidata.ucar.edu/content/support
****************************************************************************