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

[IDV #CZL-505760]: RTMA grids in IDV

John-

> > Played around with a file of mesonet netcdf file from MADIS and was
> > having some success, but a little lost as to how to set things up
> > properly. I was having troubles getting untracked on using aliases to
> > map the netCDF parameters in my file to those used in IDV. For example,
> > the MADIS netCDF file uses Temperature_K and I appear to need to get
> > that into IDV variable TEMP. I did that through the alias tool but still
> > wasn't having luck seeing temperature. I could plot the station
> > locations, though, without trouble.
> 
> This is a bug.  Are you using the RC2 version or the 1.3b2?  When
> you edit the aliases, it should apply immediately, but it looks
> like you have to exit and restart the IDV for that to apply.
> We'll fix that before the workshop.  Also, in you post above,
> you have Temperature_K, but the names are case sensitive, so
> make sure you are using temperature_K.
> 
> Another issue that we've addressed is the fact that we munge the
> name of the netCDF variable from temperature to temperature_K to
> make it unique.  Because the IDV is generic in nature, we don't
> have standard names like TMPK, TMPC like GEMPAK so people can
> put whatever unit (or no unit) with a given variable.  To distinguish
> these, we added the _<unit> to the name in the 2.0 release. That
> is causing problems, so we are making a mod that will allow the
> aliases to work with the un-munged variable name (e.g. temperature).
> 
> Sorry you ran into these two, but thanks, since we'll fix them.

We've addressed this in the latest release and you do not need to
edit the aliases with the MADIS files.  We've modified the code to
handle munged names.

> > I'm also a bit concerned about how practical memorywise it will be to
> > bring up one of the huge RTMA grids and overlay one of these huge netCDF
> > grids. I seem to need about 1 gbyte for just the netCDF file and I
> > suspect the RTMA grid is really going to be a problem. Will the boxes in
> > the class lab have enough memory to handle that? Is there a way to
> > sectorize the grid to save memory?
> 
> You can sectorize any grid through the data source properties (1.3b2)
> and through the field selector (2.0).  You can subset on a lat/lon box
> and/or by skipping grid points.

Subsetting of the RTMA grids in the latest version works quite well.

You can get the latest version (2.0RC3) from:

http://www.unidata.ucar.edu/software/idv/release/nightly/

This is the version we will use in the workshop.

Don Murray



Ticket Details
===================
Ticket ID: CZL-505760
Department: Support IDV
Priority: High
Status: Open