NOAA / NWS's Meteorological Development Laboratory
Degrib (NDFD GRIB2 Decoder)
Subject: degrib version 0.85
Date: Wed, 20 Nov 2002 18:40:06 -0500
Organization: DOC/NOAA/NWS - National Weather Service
To: Arthur Taylor
You all have the unique distinction of being registered on my degrib
mailing list (all 5 of us) (I'd like to welcome Albion Taylor and Ron
Murphy to the list). Since you are a member, you get this email...
Please let me know if you want to unregister.
I have just finished version "0.85" (last one was "0.80" on 11/05/2002)
New to this release:
1) All the code is now properly (in my opinion) commented (awaiting
internal code walk throughs) A number of misconceptions and bugs have
been resolved during the commenting process.
2) Added command line option: -revFlt generates a .tlf file which starts
in the lower left corner instead of the upper left corner. ArcView
prefers the file to start in the upper left, while NDFD Image creation
software prefers the lower left.
3) Added command line option: -MSB (-nMSB) generates binary files Most
Significant Byte first (or not) when there is a choice. This mainly
applies to .tlf or .flt files. The default I have used is for MSB .flt
files, although ArcView's S.A. uses a helper file so it can handle
both. I chose MSB for the default, because the grib2 file was MSB.
Unfortunately for the NDFD Image creation software, it assumed the file
was generated on the same platform it was running, on which happened to
be Linux, so it needed -nMSB.
4) It has now been tested successfully on 2 AIX platforms (Suitland's
mainframe, and on the ARL's AIX (courtesy of Albion Taylor.))
5) It can handle the "new" NDFD grib2 file format as well as the old.
Most of the test suites use the old format, but I do have 1 test with
the new format. The old format was a 21 byte wmo descriptor followed by
1 GRIB message after another, and only one variable was in a file.
The new format has 40 bytes (19 byte size descriptor and a 21 byte wmo
descriptor) to describe a group, then 40 bytes to describe a GRIB
message, the GRIB message, another 40 bytes to describe the next GRIB
message, the next GRIB message, etc, for all GRIB messages in that
group. (A group is 1 particular weather element, for example: Tmax).
The file now can have multiple weather elements (or groups). So it
could look like: 40 40 GRIB 40 GRIB 40 GRIB 40 40 GRIB 40 GRIB etc.
(Note if you use the new test, there was something wrong with the
creation of the .grb file, so for the second group it was only using 18
bytes (not 21) for the wmo descriptors.)
What this means to you... You are using tcldegrib, and again not much
has changed, other than bug fixes. However you may want the newer test
suite (11/15 for Tmax/ Tmin/ T/ Td/ Pop12 instead of 10/17). Also you
may want the newer code in case you get some of the "new format" data.
I think they are going to be putting some of the regions on the ftp site
soon, and you may need the new code to read it.
What this means to you... I again have compiled current versions in
~taylor/degrib/bin on rainbow (HP-GDP), in ~taylor/degrib/bin on bd10
(linux), and in /nfsuser/g06/we25at/degrib/bin on Suitland's IBM. So
again, you shouldn't need to download the code or compile it yourself.
Not sure if you are using it to help with the verification stuff or not,
but let me know if you need help.
What this means to you... Not much, although it might be nice if you
could test the current version on your system. I have tested the
current one on Suitland's AIX system.
What this means to you... I have no idea, since I don't know how you are
using it (or if you are). Hopefully I have gotten rid of more bugs than
I introduced during all the commenting, and made it more legible. So if
you are looking at the code you might want the newer copy. Also you may
need it for the "newer" ndfd grib2 files.