Re: Re[8]: 19990120: Help with decoders!

• To: Jason Burks <Jason.Burks@xxxxxxxx>
• Subject: Re: Re[8]: 19990120: Help with decoders!
• From: Robb Kambic <rkambic@xxxxxxxxxxxxxxxx>
• Date: Tue, 2 Feb 1999 14:15:12 -0700 (MST)

On 2 Feb 1999, Jason Burks wrote:

     Robb,

I have found out a little more about the NOGAPS it is on the 240 grid. Now what do I have to modify to make gribdump to recognize the new grid? The files I have noticed that I think I have to change are

Jason,

Here's the changes I would make, I'll probably miss some changes so you
will have to use the debugger.

In gbytem.c, under case FNOC add, with ? the size of the grid in bytes:

case  240: return full_gbytem(?);

     models.h, add:

#define MODEL_240  ...

statement.


     models.c

A new switch statement needs to be added for CENTER_FNOC. ie

   case CENTER_FNOC:
       switch(model) {         /* does anyone know names for these? */
       case MODEL_240:
           return "description here";

...

parameters, nx=,ny=, la1=, dx=,etc).

Yes, add under case CENTER_FNOC: after 223 ....

case  240: return fnoc_240();

Then create procedure fnoc_240 using one of the other models as a
template. You will have to look at the 240 structure and the gdes.h file
for this part.

This is going to a trial and error method of building a completely new
grid. Be patience, it's not simple because of the numerous possibilities.
The ruc2.cdl file took me 5 days to complete and I had experience building
complex cdls.  After you complete this, could I have the changes to enter
into the code in case some one else wants a FNOC model?

Good luck

Thanks,

then I will be on my way. If you could just tell me how to add a new model that would be great: Like what files I need to edit, etc.

     thanks once again
     Jason


______________________________ Reply Separator _________________________________
Subject: Re: Re[6]: 19990120: Help with decoders!
Author:  rkambic@xxxxxxxxxxxxxxxx at EXTERNAL
Date:    1/29/1999 3:44 PM


On 29 Jan 1999, Jason Burks wrote:

>

>      Robb,
>      Sorry about all of the questions

> I am now trying to decode the Navy's NOGAPS model data supplied in > GRIB into netcdf format.

Jason,

Contact Navy's NOGAPS people to see what they use to read there own grib files. Also try to find out what is the model structure of NOGAPS. I believe that gribtonc will have to be modified to translate these type of grids. When I tried to use gribdump -h or grib -b, I > get nothing. When I use gribdump -v on the file I get messages like > the following saying:

>      oversize GRIB product
>      error reading GRIB product
>      It exits saying 0 WMO Msgs, 0 GRIBs decoded, 0 written
>      is this because the the nogaps is not described anywhere in the
>      program? What do I need to do to get it to work? The data is on a

> standard grid, I think 207. Thanks for all of your help, and sorry to > take your time. In file gdes.c at 1356 there is a definition for grid 207, does this match the params of the raw file. At this point a debugger would surely pinpoint the problem. The gribtonc program is tailored to FOS/NOAAport types of products, it's not a general translator. Here's a url to another grib reader, maybe it will work better: http://wesley.wwb.noaa.gov/wgrib.html Robb... > Jason

>

=============================================================================== Robb Kambic Unidata Program Center

Software Engineer III                  Univ. Corp for Atmospheric
 Research

rkambic@xxxxxxxxxxxxxxxx WWW: http://www.unidata.ucar.edu/ ==============================================================================

==============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
rkambic@xxxxxxxxxxxxxxxx                   WWW: http://www.unidata.ucar.edu/
==============================================================================