20030407: Some information about IDV that I hope may benefit ... (or FYI)...from John McCormick..

>From: address@hidden
>Organization: EPA
>Keywords: 200304071857.h37Ivn7U019251 IDV netCDF files

John-

>Don: If you already know that IDV cannot access these types
>of NETCdf files, then just ignore this message ...   But if any one
>if these files are supposed to be "accessible" and "interpretable"
>by IDV than let me know... (There might be something incorrect in
>the way that I might be trying to "open" them)...
>
>The following files were not accessible by IDV:
>
>
>    (See attached file: 03030920_buoy.nc)    (See attached file:
>20030318_USData_test_CO.nc)   (See attached file: model_rossby_new.cdf)
>(See attached file: 20030318_USData_test_rh.nc)
>
>
>(Comment:   I did notice that the first three bytes of NETCdf files
>turned out to be "CDF" ... which to me
>  means that the file is supposedly a NETCDF file.... I also understand
>that as Stuart had mentioned to
>me that not all NETCDF files are "readable" by IDV... But these enclosed
>files are only "FYI", or in case
>there's something about them that are supposed to be "readable" or
>"interpretable" by IDV........ Thanks),

netCDF files must conform to certain conventions to be able to
be read into the IDV.  Unfortunately, many netCDF files do not
adhere to these.  For the particular files you sent, here
is what I found:

03030100_metar.nc - The IDV handles this type of file if there
                    is sufficient metadata included.  In this
                    particular file, there are is no lat/lon info
                    for any of the reports.

03030920_buoy.nc -  this is similar to the metar files and
                    will be supported in the next release.  For
                    these types of files, we rely on the naming
                    convention (not a good method, but it works
                    until we get something better).  I've added
                    .*_buoy.nc to the list of files read by the
                    NetcdfMetarDatasource and was able to plot
                    this file.

20030318_USData_test_rh.nc - I was able to read this file in okay
                             and plot contours.  Was there a 
                             particular display you couldn't get
                             working?  One note, however is that
                             the units for RHUM are invalid. 
                             ("fraction[0-1]"). 

20030318_USData_test_CO.nc - In this one, the vertical level unit
                             is invalid ("ln(p0/p)").  The IDV
                             can now can only handle vertical
                             units of pressure and altitude.

model_rossby_new.cdf - This is a cross section line or something
                       other than a 3D grid.  In what way would you 
                       want to display such a file?


We are always looking to improve the number and scope of
netCDF files that we can handle and examples like this are
useful.  Thanks for passing these along.

Don Murray