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

20000421: XCD grid decoding problems at UVa (cont.)

>From: Anthony James Wimmers <address@hidden>
>Organization: UVa
>Keywords: 200004091534.JAA08082 XCD DMGRID ETA MRF FTP backup RTMODELS.CFG 
>NOGRIB.CFG

Tony,

Last week, you noted:

>Since about last Thursday (the day we had the long phone call and
>did the buffer overhaul) the MRFs came in for about one day, but 
>ever since then, we've only gotten the MRFs >72HR forecasts:

>      131             2000111       5000   ALL  00Z MRF 72 HR< VT<=96 HR     
>      141             2000111       5000   ALL  00Z MRF > 96  HR             

and I noted that the UPC and CU/CIRES were getting all of the MRF data.

Then I said that I would login to windfall and poke around to see if
I could find a problem.  Well, I did on Saturday, and I did make a change:

I edited ~ldma/etc/ldmd.conf and change the request line that was in place
to requesting everything:

was:

request        DDPLUS|MCIDAS|HRS
       "(^[A-OQ-X])|(^[YZ].[^AHIJRU])"
       navier.meteo.psu.edu

is now:

# Test change of request pattern by Yoksas on 20000222.1003
request DDPLUS|MCIDAS|HRS ".*" navier.meteo.psu.edu

I just checked and see that the MRF stuff is now all there:

GRDLIST RTGRIDS/ALL.ALL FORM=FILE
DATASET NAME: RTGRIDS/ALL
 Dataset Position  Creation Date Max Grids         Directory Title
 ----------------  ------------- --------- -------------------------------
        5             2000115       5000   Misc. Grids for 24 APR 2000
       54             2000114       5000   ALL  00Z NGM 0  HR<=VT<=24 HR
       55             2000115       5000   ALL  00Z NGM 0  HR<=VT<=24 HR
       64             2000114       5000   ALL  00Z NGM 24 HR< VT<=48 HR
       65             2000115       5000   ALL  00Z NGM 24 HR< VT<=48 HR
       74             2000114       5000   ALL  12Z NGM 0  HR<=VT<=24 HR
       84             2000114       5000   ALL  12Z NGM 24 HR< VT<=48 HR
      104             2000114       5000   ALL  00Z MRF 0  HR<=VT<=24 HR
      105             2000115       5000   ALL  00Z MRF 0  HR<=VT<=24 HR
      114             2000114       5000   ALL  00Z MRF 24 HR< VT<=48 HR
      115             2000115       5000   ALL  00Z MRF 24 HR< VT<=48 HR
      124             2000114       5000   ALL  00Z MRF 48 HR< VT<=72 HR
      125             2000115       5000   ALL  00Z MRF 48 HR< VT<=72 HR
      134             2000114       5000   ALL  00Z MRF 72 HR< VT<=96 HR
      135             2000115       5000   ALL  00Z MRF 72 HR< VT<=96 HR
      144             2000114       5000   ALL  00Z MRF > 96  HR
      145             2000115       5000   ALL  00Z MRF > 96  HR
 ...

So, it is most likely that the previous request pattern that was being
used was excluding those grib messages from being sent to your
machine.  If you decide to stay with the new request (which is
"everything") then you will need to modify ldmd.navier.meteo.psu.edu
and ldmd.blueskies.sprl.umich.edu.

Tom

>From address@hidden Mon Apr 24 14:57:27 2000
>To: "Jennie L. Moody" <address@hidden>
>In-reply-to: Your message of "Mon, 24 Apr 2000 15:08:35 -0400."
>Date: Mon, 24 Apr 2000 14:57:19 -0600

Jennie,

>We are still having problems with not seeing the mrf grids
>we expect to see here....Tony is going to call you about this,
>sorry about that, but it is a last resort kind of situation.

Tony called.  I believe that you are getting and decoding everything.
The difference is that the grids formerly labeled as MRFs are now
labeled as AVNs.  I told Tony to check to see if the AVN GRID files
did have the fields that you need, and to get back with me if
you are not.  I noted that you (UVa), the UPC and CU/CIRES (Hsie)
are now producing the exact same set of output GRID files from XCD,
so I feel like things are working correctly (or all are not working
in exactly the same way).  I _really_ need to know if you are
not decoding the grids you need since this would mean that there
is some sort of systematic problem in DMGRID and/or McIDAS-XCD
grid decoding configuration files.

>I feel like we are supposed to be catching things and we
>are more of a data seive at the moment.... stuff is slipping
>through, and its always more of a problem to retrieve later.

I agree that going back and getting the data is a big pain.  I
don't think that you will be needing to do that, however.

I trust that you saw my earlier message to the effect that I
changed the request line in your ~ldma/etc/ldmd.conf file.  This,
I believe, is what allowed all of the MRF forecast times to
be transferred from your upstream site so they could be decoded.

>We are in another mission (started yesterday, ends next
>week), so the heat it on again....

OK.  Got it.

>Sorry to be the source of ongoing questions.  But, I hope
>you can help Tony.

I can and will.  Please have Tony contact me if he finds that you are
missing grids.

>From: "Jennie L. Moody" <address@hidden>
>Date: Mon, 24 Apr 2000 17:03:01 -0400 (EDT)

>Tom,

>Thanks.  I just got off the phone with Tony.  He thinks
>everything is okay.  We will let the next set of grids
>come in, and let our batches run, and see if this has
>straightened everything out.  I appreciate your help.

>Jennie

>From: "Jennie L. Moody" <address@hidden>
>Date: Wed, 26 Apr 2000 20:29:28 -0400 (EDT)
>Subject: mrf grids

re: I _really_ need to know if you are not decoding the grids ...

>Yes, it does appear that we are getting everything correctly.
>We always got the AVN grids in the past, but they did not get
>broken out and labelled as AVN's.  Was this a change that
>resulted when you change the RTMODELS.CFG file, or was it
>somehow related to the change in the ldmd.conf ALLOW line
>change to allow everything?  I suppose it doesn't really
>matter, but I am not clear why things broke.  My guess is that
>if one doesn't specifically file the AVN grids seperately,
>they automatically get decoded into the MRF grid designation,
>and this is what we were doing before.

re: the request line in your ~ldma/etc/ldmd.conf file

>Right, although before we made the other changes, or before
>the problems attendent to the eta grid temporal extension,
>we had been getting the MRF initializations, the AVN run
>of the MRF and the MRF forecasts.

re: have Tony contact me if he finds that you are missing grids

>Tony updated me right after he spoke with you.  Sorry I
>am just getting around to replying, but thought I would 
>wait long enough to see that things are stable.  All
>appears aok.   Once again, many thanks are due.