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

[IDV #RTX-619030]: problem with smoothing grid with lambert conformal projection

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [IDV #RTX-619030]: problem with smoothing grid with lambert conformal projection
  • Date: Wed, 29 Dec 2010 14:20:21 -0700
Hi John-

Don Murray here.  They brought in the big guns. ;-)

> > Full Name: john horel
> > Email Address: address@hidden
> > Organization: uu
> > Package Version: 2.9u2 build date:2010-12-21 08:14 UTC
> > Operating System: Windows 7
> > Hardware: Java: home: C:\Program Files\IDV_2.9u2\jre version: 1.6.0_20 
> > j3d:1.5.2 fcs (build4)
> > Description of problem: applySmoothing
> > Gridded2DSet: samples value #0 may not be missing
> >
> > So we generate grids that don't look so hot in IDV mainly because we can't 
> > smooth them. They are available via our catalog and should be in the bundle 
> > here too.
> >
> > http://wx3.chpc.utah.edu:8080/thredds/catalog/pcaps_data/UU2DVAR/20101226/catalog.html?dataset=root%20data%20directory/UU2DVAR/20101226/2010122606_2dvar.nc
> >
> > They appear fine in ncview. The fill value in the netcdf file is set 
> > properly. And, when we sectorize to zoom into a smaller area where there 
> > shouldn't be any missing values, then the projection issue causes problems.

The problem is that there are many missing lat/lon values in the file.  As Yuan 
points out, the standard CF projection parameters are not defined 
(grid_mapping, etc), so we have to use the lat/lon values we're given.  The 
initial grid reading code handles the missing values (even though in hindsight, 
it shouldn't), but the smoothing code does not.

> > Any ideas how to either zoom into an area where there won't be a problem to 
> > smooth, or handle this from the outset?

I made some changes that will get around the smoothing problem you are facing.  
 However, I suggest you look at the code that generates the netCDF files to:

- have not missing lat/lon values
- adhere to the CF conventions documented here:

http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.5/cf-conventions.html

The netCDF or THREDDS groups at Unidata can help you to make your files CF 
compliant.  In the end, that will benefit anyone who tries to use your data.

The fix will be in the next nightly build (12/30).

Don Murray


Ticket Details
===================
Ticket ID: RTX-619030
Department: Support IDV
Priority: Normal
Status: Open