20040517: netCDF cdl for irregular observation values

Stuart,

>Date: Sat, 15 May 2004 20:22:45 -0600
>From: Stuart Wier <address@hidden>
>Organization: Unavco
>To: Steve Emmerson <address@hidden>
>Subject: Re: 20040514: netCDF cdl for irregular observation values
>Keywords: 200405141955.i4EJtZIK019172 netCDF

The above message contained the following:

> Ok that is clear but I also want lat lon and level to be netcdf 
> coordinates so that they are recognized as coordinate, as in the IDV.
> 
> So what we have is a netcdf file holding a list of locations (lat, lon, 
> level) and a variable or several variables at each location - and the
> lat lon and level values are netcdf coordinates too.  It appears to me 
> that if you give a netcdf file a list of lat and lon coordinates then
> a grid with them as the edges is assumed, and this is not a grid.

You've correctly identified the problem: the netCDF data model can't
handle an irregular grid.  I'm afraid that there's no way to do what you
want within the netCDF data model.

One could do something like import the dataset into VisAD (as an
IrregularGrid) and then use the facilities of that package to derive the
independent variables on a regular grid -- but stuff like that has to be
done outside the netCDF package.

Regards,
Steve Emmerson