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

20030829: McIDAS ADDE definition of RTGRIDS/MRF

>From: "Fingerhut, William A" <address@hidden>
>Organization: Lyndon State
>Keywords: 200308291641.h7TGfNLd017978 McIDA-XCD MRF AVN GFS RTMODELS.CFG

Hi Bill,

>I've been checking out our mcidas scripts on the latest version of
>mcidas and found something strange.
>
>A script was failing on a GRDCOPY command, so I experimented with a
>similar GRDLIST to check data availability on our server, omega.
>
>mcuser@omega:~/mcidas/data> grdlist.k RTGRIDS/MRF PARAM=Z LEV=500
>FRANGE=144 240 12 DAY=2003241 TIME=0 GPRO=MERC NUM=99
>grdlist.k: No grid found matching search conditions
>GRDLIST - done

The MRF model data, which was actually the initialization of the AVN
model is gone since all of the data is available in the AVN, which
was renamed GFS.  Confusing right!?  Thank the weather service.

>If I lower the starting hour in the FRANGE keyword: (data is present)
>
>mcuser@omega:~/mcidas/data> grdlist.k RTGRIDS/MRF PARAM=Z LEV=500
>FRANGE=108 240 12 DAY=2003241 TIME=0 GPRO=MERC NUM=99
>Dataset position 41     Directory Title= ALL  00Z GFS 96 HR< =
>VT<=120HR
>PAR    LEVEL       DAY        TIME    SRC FHOUR     FDAY       FTIME GRID   PR
> O
>---- --------- ------------ -------- ---- ----- ------------ -------- ----- --
> --
>Z     500 MB   29 AUG 03241 00:00:00  GFS   132 03 SEP 03246 12:00:00 615 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   144 04 SEP 03247 00:00:00 620 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   120 03 SEP 03246 00:00:00 625 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   108 02 SEP 03245 12:00:00 630 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   156 04 SEP 03247 12:00:00 633 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   180 05 SEP 03248 12:00:00 638 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   168 05 SEP 03248 00:00:00 642 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   192 06 SEP 03249 00:00:00 647 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   204 06 SEP 03249 12:00:00 651 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   216 07 SEP 03250 00:00:00 655 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   228 07 SEP 03250 12:00:00 659 MERC
>Z     500 MB   29 AUG 03241 00:00:00  GFS   240 08 SEP 03251 00:00:00 664 MERC
>Number of grids listed = 12
>GRDLIST - done

I would say that your setup for RTGRIDS/MRF should be deleted/not used
and you should use RTGRIDS/GFS, the new definition.  This will require
that your ADDE dataset definitions be up to date, and that the entries
in ~mcidas/workdata/RTMODELS.CFG be current (as per v2003 installation
instructions).

>I get the same results on freeBSD and linux pc's.
>
>If I point to motherlode, there is no dataset RTGRIDS/MRF . Who is right?

There should be no RTGRIDS/MRF dataset now that the data is gone.  I
left in the definition for the dataset group/descriptor, but I am
not populating it with data.

>If I use RTGRIDS/GFS on motherlode, I reproduce the same strange
>behavior - unexpected results with FRANGE=108 ...   .

The second part of your observation, that the first positional value
of the FRANGE= keyword has to be less than 128 in order to work
is demonstrating a bug that I had not known about until you pointed it
out.

>So, something is not right here ?

You are right that having to specify FRANGE=108 240 instead of FRANGE=144 240
appears to be a bug, and I will have to dig into the code to figure out
what is going on.

The RTGRIDS/MRF dataset should be viewed as being archaic.  The RTGRIDS/GFS
dataset is defined to be the exact same as RTGRIDS/AVN on machines running
McIDAS-XCD v2003 and that are setup correctly.  If your machine(s) are
not setup correctly, I can jump in to help them get setup correctly.

>What do you think ? Bill

Thanks for reporting the FRANGE= bug of GRDLIST.  I will look into it.

Tom

>From address@hidden Fri Aug 29 11:57:46 2003

Tom,

The problem is also in GRDCOPY, and maybe other places ?

Thanks for everything. Next time we need help we'll send
you a call for help.

Bill