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

20011020: New NOAAPort GRIB.

Stonie,

The grid you are refering to is:
ZDIA98 KKCI
Broadcast every 5 minutes for National convective weather forecast.
See the ADMIN message below regarding its availability.

The one thing to watch when changing LLMXGD is the heap that
is also defined in theinclude file. Generally, you 
would have to have a heap at least 3x the size of the grid
to do any grid computations on the grid. But just to display it,
that isn't going to be the limitation.

Most OSs have a maximum user stack size. If you compile a program
with size greater than that, you will get core dumps. In C, you
can do a malloc() rather than statically define the array, and
then you won't be using the program stack....but if Fortran,
you are statically defining the array- and that effects the program
size on the stack.

There is also the 4KM precipitation grid. I don't think
I will be redefining the default grid size bigger than 400,000
in the 5.6e release. Perhaps I will update dcgrib2 to check the grid size
and compare it against the defined LLMXGD size and gracefully handle the
situation where the prod is greater than allowed.

Steve Chiswell
Unidata User Support

424 
NOUS41 KWBC 262055
PNSWSH

PUBLIC INFO STATEMENT...TECH. IMPLEMENTATION NOTICE 01-23 AMENDED
NATIONAL WEATHER SERVICE HEADQUARTERS WASHINGTON DC
500 PM EDT WED SEP 26 2001

TO:       FAMILY OF SERVICES /FOS/ SUBSCRIBERS...NOAA WEATHER
          WIRE SERVICE /NWWS/ SUBSCRIBERS...EMERGENCY MANAGERS
          WEATHER INFORMATION NETWORK /EMWIN/ SUBSCRIBERS...OTHER
          NATIONAL WEATHER SERVICE /NWS/ CUSTOMERS OF AVIATION
          FORECASTS...NWS EMPLOYEES

FROM:     MARK ANDREWS
          CHIEF...AVIATION SERVICES BRANCH

SUBJECT:  SECOND AMENDMENT - IMPLEMENTATION OF THE NATIONAL
          CONVECTIVE WEATHER FORECAST PRODUCTS...EFFECTIVE
          SEPTEMBER 27 2001

...THE FOLLOWING CHANGES HAVE NO IMPACT ON NOAA WEATHER WIRE
SERVICE SUBSCRIBERS...

...THIS SUPERCEDES TECHNICAL IMPLEMENTATION NOTICE 01-23...
ISSUED AUGUST 16 2001...AND THE FIRST AMENDMENT TO 
TIN 01-23...ISSUED AT 1655 UTC TODAY...

EFFECTIVE THURSDAY SEPTEMBER 27 2001 AT 1300 UNIVERSAL
COORDINATED TIME /UTC/...THE NWS AVIATION WEATHER CENTER IN
KANSAS CITY MISSOURI WILL BEGIN DISSEMINATING THE NATIONAL
CONVECTIVE WEATHER FORECAST /NCWF/.  THE NCWF CONSISTS OF TWO
PARTS ...A DETECTION PRODUCT AND A ONE HOUR FORECAST.  

1./  THE DETECTION PRODUCT IS A GRID OF CONVECTIVE AVIATION
HAZARDS DISTRIBUTED IN GRIB FORMAT.  THE WMO HEADING OF THE GRIB
DETECTION PRODUCT HAS CHANGED FROM ZDIZ98 KKCI TO ZDIA98 KKCI.
THE GRIB DETECTION PRODUCT IS GENERATED AT FIVE MINUTE INTERVALS
AND IT IS AVAILABLE VIA ANONYMOUS FTP ON THE NWS FTP SERVERS. 
AVAILABILITY OF THE GRIB PRODUCT ON THE NOAAPORT SATELLITE
BROADCAST NETWORK /SBN/ IS PLANNED FOR OCT 22 2001.
 
THE WMO HEADING FOR THE GRIB DETECTION PRODUCT IS:
     WMO HEADING:  ZDIA98 KKCI
                      |
      CORRECTED FROM TIN01-23 DATED 8/16/01

2./  THE ONE HOUR FORECAST PRODUCT IS A COLLECTION OF POLYGONS
DELINEATING FORECAST POSITIONS OF CONVECTIVE AVIATION HAZARDS AND
IS DISTRIBUTED IN BUFR FORMAT.  THE BUFR FORECAST PRODUCT IS
GENERATED AT FIVE MINUTE INTERVALS AND IS AVAILABLE VIA ANONYMOUS
FTP ON THE NWS FTP SERVERS.  THE BUFR FORECAST PRODUCT IS ALSO 
AVAILABLE NOW ON THE NOAAPORT SBN.  ITS WMO HEADING IS UNCHANGED
AND IS JSAT98 KKCI.

OPERATIONAL IMPLEMENTATION OF THE NCWF IS SCHEDULED FOR THURSDAY
SEPTEMBER 27 2001 AT 1300 UTC.

IF YOU HAVE ANY QUESTIONS REGARDING THESE NEW PRODUCTS... PLEASE
CONTACT:

     MARK ANDREWS
     CHIEF...AVIATION SERVICES BRANCH
     PHONE:  301-713-1726 X 109
     address@hidden

THIS AND OTHER NWS TECHNICAL IMPLEMENTATION NOTICES ARE
AVAILABLE ON THE INTERNET AT /USE LOWER CASE/:

      HTTP://WWW.NWS.NOAA.GOV/OM/NOTIF.HTM

END

>From: "Stonie R. Cooper" <address@hidden>
>Organization: Planetary Data, Incorporated
>Keywords: 200110201959.f9KJxl100163

>
>--------------Boundary-00=_HXQIHFHQOOX1XJNCEHLQ
>Content-Type: text/plain;
>  charset="iso-8859-1"
>Content-Transfer-Encoding: 8bit
>Subject: 
>
>Steve,
>
>I've been on other things lately, so I don't know how long this new GRIB 
>message has been available (see attached), but it deals some complication to 
>GEMPAK.
>
>It's a whopper of a grid - 1830 x 918 - and I don't see the parm value 
>defined in any documentation - but by it's timeliness (it comes out every 5 
>minutes) and comparing it with other charts - it seems to be a radar related 
>grid.
>
>Knee jerk was to bump up the LLMXGD to 1750000 from the 400000 it was; 
>everything compiled except the gdvint - this is with GNU g77 - complains that 
>the array is too large:
>
>vicmn.cmn: In subroutine `vg_clos':
>vicmn.cmn:54:
>        +                  rlnp (MXGD, 2*MAXLVL),
>                           ^
>Array `rlnp' at (^) is too large to handle.
>
>That was the only "make" complaint; I haven't ran everything to see what I 
>broke, but I did do a make distclean after I edited the LLMXGD and remade 
>everything.
>
>This may not be the best way to go about tackling this large GRIB type . . . 
>I just wanted to let you know about the event - in case you were not already 
>exposed to it.  I made up an entry for the grid tables so I could get it 
>decoded, and I have done a gdcntr - so far so good.  Let me know if you 
>potential pitfalls to this approach (besides breaking gdvint).
>-- 
>Stonie R. Cooper,
>Science Officer
>Planetary Data, Incorporated
>3495 Liberty Road
>Villa Rica, Georgia  30180
>ph. (770) 456-0700; pg. (888) 974-5017; fx. (770) 459-0016
>
>