[Xastir] NEXRAD decoder?

Gerry Creager N5JXS gerry.creager at tamu.edu
Sun Jan 18 11:01:56 EST 2004


Grmmmph!  Folks keep dissin' FORTrAN.  OK, I got that outta my system...

Curt Mills wrote:
> On Sun, 18 Jan 2004, Tyler Allison wrote:
> 
> 
>>Anybody know of a piece of software that will decode the NEXRAD data that
>>is available from NOAA via ftp?  Searches on google only turn up
>>commercial products or the fortran source code from the NWS. (I dont want
>>a full blown piece of weather software, nor do I know fortran)

I do create a rectilinear (non-projected flat product for the 
NM/Texas/OK area, as well as a US product.  Both of these can be 
obtained via HTTP, and are NEXRAD Level III composites.

Take a look at http://page4.tamu.edu/USrad.geo for the US product geo, 
which will work with xastir.

It's not unreasonable for me to consider additional products if folks 
have specific areas they want/need... I'm considering gridding the 
conterminus US into 6-8 sections with some overlap and doing products on 
those on a 6 min. basis.  Mostly depends on server load.

I use a package called GEMPAK written by the UCAR Unidata program.  It's 
one of the standards in academia for this sort of thing, and does a good 
job.  I just spent a week in Seattle occasionally hoisting a brew (Pike 
Brewery) with the guys who develop and maintain GemPak.  Asked a lot of 
questions, learned a lot, gave feedback, taught a little.

In general, the data you're talking about is referred to as Level III 
NEXRAD data.  The reference site for most of the work on NEXRAD is here: 
http://www.nssl.noaa.gov/special.shtml

GemPak is distributed to appropriate .edu's and may be distributed if 
you can make a case to Unidata as to why you're interested.  Source is 
available, although it's my considered opinion, and I griped about this 
last week, that using their binaries is easier and more reliable than a 
source installation.  A CAVEAT IF YOU TRY INSTALLING THIS STUFF (sorry 
for shouting: it's important!):  Suspend all non-autonomic brain 
function and just follow the directions explicitly.  Do not deviate, 
even if it appears they don't make sense.  Don't read the directions 
ahead of time, or you'll be tempted to take a short-cut or do something 
the instructions tell you not to do.  Read, Install, DO NOT THINK. 
Trust me.  I've burned this bridge more'n once, and I'm an "experienced" 
user...

Once you have the software, where do you propose to get the data?  There 
are data feeds available, but the commercial ones aren't cheap and the 
free ones are generally reserved to .edu's again.  Once more, there may 
be ways around this.  It's something we can look into.  I get a feed or 
3 here (I've got 4 machines up getting different feeds in, including a 2 
terabyte server that archives *all* the wx data for up to 2 months 
on-line) but I'll have to check acceptable-use/distro requirements on radar.

If you're looking for a project, the Level II data is coming on-line and 
a powerful contribution to the community could be made by someone 
looking to develop new code to work with that.  It's near-real time 
(current latencies are on the order of 3 sec from radar detection) and 
there is considerably less on-site processing of these before they leave 
the radar site than on the older, more ubiquitous Level III data. 
Again, look at the NSSL site above for more info.

Level II data is on the order of a GB/hr for the current full feed 
across my big site.  Level III (NNEXRAD) NNEXRAD looks like about 82-83 
MB/hr.  Check out my site (bigbird) at 
http://my.unidata.ucar.edu/content/software/idd/rtstats/siteindex.php?bigbird.tamushsc.edu 
for more info on feeds, latencies, etc.

> Well, my first "official" computer class was in fortran, although I
> was doing hand-coded machine code before that.
> 
> I know there's a translator from Pascal to C, and I think I remember
> something about one to translate fortran, so it might not be that
> hard to turn it into something that's a bit more modern.
> 
> Did "apropos fortran" on this system and got this:
> 
> 
> f2c (1)              - Convert Fortran 77 to C or C++ . . .
> g77 (1)              - GNU project Fortran 77 compiler
> libfpvm (3)          - PVM C and Fortran programming libraries
> libpvm3.a (3) [libfpvm] - PVM C and Fortran programming libraries
> libpvm (3)           - PVM C and Fortran programming libraries
> 
> 
> The pascal translator is p2c.  There's are multiple fortran and
> pascal compilers also if you don't want to translate to C.
> 
> 
> 
>>I simply want a way to decode the NEXRAD file and turn it into some
>>graphics format (gif, jpg, png, whatever)

GemPak or Unisys's WxP.  There are several other codebits out there, but 
those are the 2 biggies.  Be forwarned, WxP, as a commercial product is 
expensive.

> Sorry, don't know anything about the format myself.  Just heard the
> term before.
> 

There will be a test next class period...
-- 
Gerry Creager -- gerry.creager at tamu.edu
Texas Mesonet -- AATLT, Texas A&M University	
Cell: 979.229.5301 Office: 979.458.4020 FAX: 979.847.8578
Page: 979.228.0173
Office: 903A Eller Bldg, TAMU, College Station, TX 77843




More information about the Xastir mailing list