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

20000420: XCD grid decoding problems at UVa



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

Tony,

>This may not be a problem, just a misunderstanding that you'll have to
>clear up for us. I tried to make my message long to cut down on your work.
>;)
>
>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:
>
>GRDLIST RTGRIDS/ALL.ALL FORM=FILE                                             
>DATASET NAME: RTGRIDS/ALL                                                     
> Dataset Position  Creation Date Max Grids         Directory Title            
> ----------------  ------------- --------- -------------------------------    
>        1             2000111       5000   Misc. Grids for 20 APR 2000        
>       51             2000111       5000   ALL  00Z NGM 0  HR<=VT<=24 HR      
>       61             2000111       5000   ALL  00Z NGM 24 HR< VT<=48 HR      
>      131             2000111       5000   ALL  00Z MRF 72 HR< VT<=96 HR      
>      141             2000111       5000   ALL  00Z MRF > 96  HR              
>      201             2000111       5000   ALL  00Z MAPS                      
>      211             2000111       5000   ALL  03Z MAPS                      
>      221             2000111       5000   ALL  06Z MAPS                      
>      231             2000111       5000   ALL  09Z MAPS                      
>      241             2000111       5000   ALL  12Z MAPS                      
>      401             2000111       5000   ALL  00Z AVN 0  HR<=VT<=24 HR      
>      410             2000110       5000   ALL  00Z AVN 0  HR<=VT<=24 HR      
>      411             2000111       5000   ALL  00Z AVN 24 HR< VT<=48 HR      
>      420             2000110       5000   ALL  00Z AVN 24 HR< VT<=48 HR      
>      421             2000111       5000   ALL  00Z AVN 48 HR< VT<=72 HR      
>      430             2000110       5000   ALL  00Z AVN 48 HR< VT<=72 HR      
>      440             2000110       5000   ALL  12Z AVN 0  HR<=VT<=24 HR      
>      450             2000110       5000   ALL  12Z AVN 24 HR< VT<=48 HR      
>      460             2000110       5000   ALL  12Z AVN 48 HR< VT<=72 HR      
>      501             2000111       5000   ALL  00Z ETA 0  HR<=VT<=24 HR      
>      511             2000111       5000   ALL  00Z ETA 24 HR< VT<=48 HR      
>      521             2000111       5000   ALL  00Z ETA > 48  HR              
>      531             2000111       5000   ALL  12Z ETA 0  HR<=VT<=24 HR      
>      541             2000111       5000   ALL  12Z ETA 24 HR< VT<=48 HR      
>GRDLIST - done                                                                

This has not been my or the other site that I have been working with
experience.

>But we're not instructing it to lose the other MRFs:
>
>SCRATCH=5001
>NGM=  3 5051 120000 240000 480000
>MRF=  1 5101 120000 240000 960000
>MAPS= 0 5201 30000
>MDR=  2 5301
>AVN=  3 5401 120000 240000 720000
>ETA=  1 5501 120000 240000 480000
>
>(From RTMODELS.CFG)

OK.  You should be saving all of the MRFs.  The fact that you were
for awhile tells us that there is probably another problem.

>Now I see that the exact format of grids appear in the AVN:
>
>GRDLIST RTGRIDS/AVN.1 NUM=ALL                                                 
>Dataset position 1      Directory Title= ALL  00Z AVN 0  HR<=VT<=24 HR        
>        
>PAR    LEVEL       DAY        TIME    SRC FHOUR     FDAY       FTIME  GRID  PRO
>---- --------- ------------ -------- ---- ----- ------------ -------- ----- ---
>T     850 MB   20 APR 00111 00:00:00  AVN    12 20 APR 00111 12:00:00 1    MERC
>T     700 MB   20 APR 00111 00:00:00  AVN    12 20 APR 00111 12:00:00 2    MERC
>T     500 MB   20 APR 00111 00:00:00  AVN    12 20 APR 00111 12:00:00 3    MERC
>T     400 MB   20 APR 00111 00:00:00  AVN    12 20 APR 00111 12:00:00 4    MERC
>T     300 MB   20 APR 00111 00:00:00  AVN    12 20 APR 00111 12:00:00 5    MERC
>T     250 MB   20 APR 00111 00:00:00  AVN    12 20 APR 00111 12:00:00 6    MERC
>T
>
>(etc)
>
>But no SRC=MRF grids appear in this gridfile.

No, the SRC will be AVN.

>So my question is, did the
>AVN suddenly replace the MRF in the datastream?

Here is the information that described the change that was put in place
when Version 7.6 was released

Release Notes
http://www.unidata.ucar.edu/packages/mcidas/mcx/release_notes.html

Aviation Run of the MRF Model

Beginning with this upgrade, grids from the Aviation run of the Medium
Range Forecast (MRF) model will be labeled AVN. In previous versions,
these grids were labeled MRF.  The reasons for the label change are
listed below.

     The Aviation and MRF runs use different data. The MRF run is
     started after the Aviation run and uses additional data in its
     initialization grids.

     Giving different labels to the grids provides increased
     flexibility in grid filing methods.

     Most grid users differentiate between the Aviation run and the MRF
     run of the MRF model.

Required Action: 

     If you are upgrading from a previous version of McIDAS-XCD,
     complete the steps below to configure the GRIB decoder to file the
     AVN grids in their own grid file range.

   1.As the user mcidas start a McIDAS-X session. Then, in the McIDAS-X Text and
     Command Window stop the GRIB decoder. 

          Type: DECINFO SET DMGRID INACTIVE

   2.From the Unix shell, change to the ~mcidas/data directory 

          Type: cd ~mcidas/data

   3.Edit the file RTMODELS.CFG and add a line similar to the one
     listed below, which
     configures the GRIB decoder to file AVN grids in the grid file
     range 5401 through 5480. If you want to file the grids in a range
     other than the default range, then change 5401 to the beginning
     grid file number of your choice. Note that the filing format
     specified in the example below will consume 80 grid files. After
     adding the appropriate line to RTMODELS.CFG, write and quit the
     file.

      AVN= 3 5401 120000 240000 720000 


   4.Restart the GRIB decoder.

          Type: DECINFO SET DMGRID ACTIVE

     Complete the steps below to add a new ADDE real-time dataset for
     the AVN grid files configured in steps 1 through 4 above.

        1.In the existing McIDAS-X session, and run a command similar
          to the one
          listed below. If you chose a different starting grid file
          number in step 3 above, adjust the numeric range in the
          command below accordingly.

          Type: DSSERVE ADD RTGRIDS/AVN GRID 5401 5480 RT=YES "Real-Time AVN 
Grids

        2.Exit the McIDAS-X session and logoff. 

>If not, what is the
>difference, and how did we lose the MRF?

Your listing of MRF files is not what it should be.  The same listing from
Hsie (CU/CIRES/NOAA) this morning was:

 igu.k LIST 5000 6000
> GRIDFILE   PROJ  CREATED MAXGRD IDENT
> --------   ----  ------- ------ ----------------
>   5001        0  2000111  5000  Misc. Grids for 20 APR 2000
>   5051        0  2000111  5000  ALL  00Z NGM 0  HR<=VT<=24 HR
>   5061        0  2000111  5000  ALL  00Z NGM 24 HR< VT<=48 HR
>   5071        0  2000111  5000  ALL  12Z NGM 0  HR<=VT<=24 HR
>   5081        0  2000111  5000  ALL  12Z NGM 24 HR< VT<=48 HR
>   5101        0  2000111  5000  ALL  00Z MRF 0  HR<=VT<=24 HR
>   5111        0  2000111  5000  ALL  00Z MRF 24 HR< VT<=48 HR
>   5121        0  2000111  5000  ALL  00Z MRF 48 HR< VT<=72 HR
>   5131        0  2000111  5000  ALL  00Z MRF 72 HR< VT<=96 HR
>   5141        0  2000111  5000  ALL  00Z MRF > 96  HR
>   5201        0  2000111  5000  ALL  00Z MAPS
>   5211        0  2000111  5000  ALL  03Z MAPS
>   5221        0  2000111  5000  ALL  06Z MAPS
>   5231        0  2000111  5000  ALL  09Z MAPS
>   5241        0  2000111  5000  ALL  12Z MAPS
>   5251        0  2000111  5000  ALL  15Z MAPS
>   5261        0  2000111  5000  ALL  18Z MAPS
>   5301        0  2000111    24  MDR data for 20 APR 2000
>   5401        0  2000111  5000  ALL  00Z AVN 0  HR<=VT<=24 HR
>   5411        0  2000111  5000  ALL  00Z AVN 24 HR< VT<=48 HR
>   5421        0  2000111  5000  ALL  00Z AVN 48 HR< VT<=72 HR
>   5431        0  2000111  5000  ALL  12Z AVN 0  HR<=VT<=24 HR
>   5441        0  2000111  5000  ALL  12Z AVN 24 HR< VT<=48 HR
>   5451        0  2000111  5000  ALL  12Z AVN 48 HR< VT<=72 HR
>   5501        0  2000111  5000  ALL  00Z ETA 0  HR<=VT<=24 HR
>   5511        0  2000111  5000  ALL  00Z ETA 24 HR< VT<=48 HR
>   5521        0  2000111  5000  ALL  00Z ETA > 48  HR
>   5531        0  2000111  5000  ALL  06Z ETA 0  HR<=VT<=24 HR
>   5541        0  2000111  5000  ALL  06Z ETA 24 HR< VT<=48 HR
>   5561        0  2000111  5000  ALL  12Z ETA 0  HR<=VT<=24 HR
>   5571        0  2000111  5000  ALL  12Z ETA 24 HR< VT<=48 HR
>   5581        0  2000111  5000  ALL  12Z ETA > 48  HR
>   5591        0  2000111  5000  ALL  18Z ETA 0  HR<=VT<=24 HR
>   5601        0  2000111  5000  ALL  18Z ETA 24 HR< VT<=48 HR
>  ---END OF LISTING

As you can see, Hsie had 0Z MRF grids from HR+0 to HR > 96.

>Keep in mind, we had this problem
>even when the buffer size was 128MB.

OK, this is good to know.

>Also, the "full" MRF came in for
>about one day after we reset the buffer last week, then came in in this
>trucated form, but then came directly in this truncated form when we
>reset the buffer yesterday.

This is what I don't understand.  I will try to logon to your machine
to see what may be going on sometime this weekend.

Tom