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

20000308: McIDAS-XCD setup and LDM questions at Colgate (cont .)

>From: Adam Burnett <address@hidden>
>Organization: Colgate
>Keywords: 200002291726.KAA26578 McIDAS-X 7.60

Adam,

Sorry I couldn't get to this first thing, but I had a dentist appt. to
go to.

>Thought you had seen the last of me. No way.

:-(

>A couple of more questions.
>
>Am I correct that the NMC grids are now provided through xcd?

XCD decodes the model data in the NOAAPORT stream that you get via the
IDD, yes.

>I remember in
>the old days, we used to get them through the mcidas data stream.

The files in the Unidata-Wisconsin datastream were very small subsets
of the data that are available.

>The
>reason I'm asking is because I'm not showing any of the NMC grids on my
>system.

>I have a number of grids in the 5000 series but nothing in the
>classic NMC grid (like GRID0003) zone.  

You won't have the old grids, GRID0001, GRID0101.  What you should have,
however, is the GRID files from which these things were made.  There are
a couple/three of things here:

o I see that you are setup to decode all of the grids in the IDD, but I
  only see MRF, AVN, and MISC grids being decoded:

  GRDLIST RTGRIDS/ALL.ALL FORM=FILE

  We need to see why you are not getting/decoding the ETA and NGM grids.

  The old GRID0001 file that was in the UW stream was NGM grids
  from the full NGM set plus grids that were created by McIDAS ROUTE
  PostProcess BATCH invocations (e.g., 250 mb speed grids).

o After you begin decoding the NGM grids, you have two choices:

  o use only the grids that are in the NGM grid data set:

    RTGRIDS/NGM

  o recreate the old GRID0001 file using the UWGRID routine that is
    bundled with 7.60.  To simplify use of UWGRID through a cron
    job, I created a shell script, uwgrid.sh, (gets installed
    in the ~mcidas/workdata directory) that can be edited to set
    McIDAS environment variables (like all of the others we have
    been talking about) and kicked off by cron at the appropriate
    times (this the the hard thing; the appropriate time depends
    on your data ingestion capabilities/latencies) to recreate
    GRID0001.  The good thing about uwgrid.sh (UWGRID) is that it
    reads the McIDAS routing table (ROUTE.SYS) to figure out
    how to name the output file (e.g., GRID0001, GRID0002, etc.)
    AND it can run the ROUTE PostProcess BATCH IF it is specified
    in the routing table.  You should take a look at ~mcidas/workdata/uwgrid.sh
    to see about setting up this creation

  o use the Fkey menu on the NGM and AVN files directly.  This is
    now setup for you.  The Fkey menu configuration/definition file
    UNIMENU.DEF (installed in ~mcidas/data) contains the entries/definitions
    ?MODEL1 and ?MODEL2.  These define the data sets that the Fkey
    menu will try to use when plotting North American or Global Initialization
    grids, respectively.  The definitions in UNIMENU.DEF look like:

    ?MODEL1   "RTGRIDS/NGM
    ?MODEL2   "RTGRIDS/AVN

    If you are doing things from an account other than 'mcidas', you may
    have a copy of UNIMENU.DEF in the account's mcidas/data directory.
    If this is the case, you should remove this file as it is old and
    does not contain all of the McIDAS string definitions that are needed.

To verify that you can start a McIDAS-X session and plot gridded data
from the Fkey menu, do the following:

DATALOC ADD RTGRIDS adde.unidata.ucar.edu

Then start the Fkey menu and try plotting model things.  Afterwards,
please do:

DATALOC ADD RTGRIDS gissun.colgate.edu

>I'm beginning the process of learning the new ADDE way of doing things.

OK, are/have you gone through the online Learning Guide:
http://www.unidata.ucar.edu/packages/mcidas/mclearn760/learn_guide.html

If not, give it a look as it deals with ADDE concepts.

>At
>the same time, I'm playing with mcidas commands via cron so that I can feed
>our department web site with weather stuff.  My initial attempt at this
>involved the creation of an eastern goes 8 IR GIF every hour.  I created a
>script entitled ir8script with the necessary (and maybe some unnecessary)
>environmental variables.  Here it is:


>PATH=/data2/mcidas/bin
>PATH=$PATH:/bin:/usr/ccs/bin:/usr/bin:/usr/ucb:/etc
>
>export PATH
>
>MCPATH=/data2/mcidas/data
>
>export MCPATH
>
>mcenv -f 700x1000 << 'EOF'
>imgdisp.k RTIMAGES/GE-IR
svgif.k 1 IR8.GIF
>exit
>EOF
>exit 0

OK.  Here is your script with my modifications:

#!/bin/sh

# Set MCHOME to be the home directory of your 'mcidas' installation.

# Set MCDATA to be the working directory of your 'mcidas' installation.
# For this example, I will assume that this is /home/mcidas/workdata
#
# Set MCPATH to include MCDATA as its first directory.  I assume
# in this example that McIDAS-X was installed in /home/mcidas.

MCHOME=/data2/mcidas
MCDATA=$MCHOME/workdata
MCPATH=${MCDATA}:$MCHOME/data:$MCHOME/help
MCGUI=$MCHOME/bin
MCLOG=$MCDATA/webcreate.log
PATH=${MCGUI}:/bin:/usr/ccs/bin:/usr/bin:/usr/ucb:/etc

# Send all textual output to the log file
exec 2>$MCLOG 1>&2

cd $MCDATA

# Now run 'mcenv' to create a McIDAS-X environment in which all of the
# scouring routines will run.
# NOTES:
#  o you need to modify the parameters in the commands to be run to
#    match your system setup.
#  o if you did not build the XCD component of the Unidata McIDAS-X
#    distribution, you should:
#      o comment out the 'qrtmdg.k GRID' entry below
#      o comment out the 'delwxt.k' entry

MCDATA=$MCDATA MCPATH=$MCPATH PATH=$PATH mcenv -f 700x1000 << 'EOF'

imgdisp.k RTIMAGES/GE-IR 1 EU=IMAGE
frmsave.k 1 IR8.GIF
exit

EOF

exit 0

>When the script runs, it tells me that the RTIMAGES/GE-IR can't be located.
>I've used this same command in mcidas and it runs fine.  Do I have to use a
>logon.k command to identify the server?  How would that look for my setup?

The problem is that the definition of MCPATH was incomplete.  Also, you
will want your script to work from the working directory for the user
'mcidas' (or whoever is running it) so it can pickup needed REDIRECTions
and the like.

Please try out my modified script and let me know if it works the way
you want it to.

>Thanks again

Later...

Tom