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

20030626: problem with GRDCOPY (cont.)

>From: Owen Cooper <address@hidden>
>Organization: Aeronomy Laboratory/NOAA
>Keywords: 200306251740.h5PHeMLd017765 McIDAS-X GRDCOPY

Owen,

>My RTGRIDS are coming from ADDE.UCAR.EDU

OK, thought so.

I found the problem.  What is happening is that the selection code is
comparing the grid number (McIDAS refers to this as the geographic ID)
between the two grids, and sees that they are different.  Given this,
it is exiting since it doesn't recognize the grids to be the same.

The following demonstrates what I mean:

1) I GRDCOPYed the U and V AVN grids for TIME=0 FHOUR=6 DAY=177
   GPRO=MERC LEV=300 to a GRID file that was empty otherwise.  The
   ADDE server used was ADDE.UCAR.EDU.Here is a long linsting of those
   grids:

GRDLIST MYDATA/GRIDS.1 NUM=ALL FORM=ALL
Dataset position 1      Directory Title=
PAR    LEVEL       DAY        TIME    SRC FHOUR     FDAY       FTIME  GRID  PRO
---- --------- ------------ -------- ---- ----- ------------ -------- ----- ----
U     300 MB   26 JUN 03177 12:00:00  AVN     6 26 JUN 03177 18:00:00     1 MERC
Total pts= 41905 Num rows= 145  Num columns= 289    received:  2003177 155813Z
u-component of wind
GRIB ID numbers: Geographic =  43; PAR = 33; Model ID = 96; Level type =100
Units of gridded variable are MPS  Scale of variable is:   2
Mercator Projection
Min Lat= -90.00  Max Lat=  90.00  Min Lon=-180.00  Max Lon= 180.00
Latitude Increment= 1.2500  Longitude Increment= 1.2500
PAR    LEVEL       DAY        TIME    SRC FHOUR     FDAY       FTIME  GRID  PRO
---- --------- ------------ -------- ---- ----- ------------ -------- ----- ----
V     300 MB   26 JUN 03177 12:00:00  AVN     6 26 JUN 03177 18:00:00     2 MERC
Total pts= 41905 Num rows= 145  Num columns= 289    received:  2003177 155825Z
v-component of wind
GRIB ID numbers: Geographic =  44; PAR = 34; Model ID = 96; Level type =100
Units of gridded variable are MPS  Scale of variable is:   2
Mercator Projection
Min Lat= -90.00  Max Lat=  90.00  Min Lon=-180.00  Max Lon= 180.00
Latitude Increment= 1.2500  Longitude Increment= 1.2500
Number of grids listed = 2
GRDLIST - done


2) next, I pointed at the SSEC RTGRIDS server and did the exact same
   GRDCOPY into a different GRID file. Here is what those grids look
   like:

GRDLIST MYDATA/GRIDS.2 NUM=2 FORM=ALL
Dataset position 2      Directory Title=
PAR    LEVEL       DAY        TIME    SRC FHOUR     FDAY       FTIME  GRID  PRO
---- --------- ------------ -------- ---- ----- ------------ -------- ----- ----
U     300 MB   26 JUN 03177 12:00:00  AVN     6 26 JUN 03177 18:00:00     1 MERC
Total pts= 41905 Num rows= 145  Num columns= 289    received:  2003177 155830Z
u-component of wind
GRIB ID numbers: Geographic =3744; PAR = 33; Model ID = 96; Level type =100
Units of gridded variable are MPS  Scale of variable is:   2
Mercator Projection
Min Lat= -90.00  Max Lat=  90.00  Min Lon=-180.00  Max Lon= 180.00
Latitude Increment= 1.2500  Longitude Increment= 1.2500
PAR    LEVEL       DAY        TIME    SRC FHOUR     FDAY       FTIME  GRID  PRO
---- --------- ------------ -------- ---- ----- ------------ -------- ----- ----
V     300 MB   26 JUN 03177 12:00:00  AVN     6 26 JUN 03177 18:00:00     2 MERC
Total pts= 41905 Num rows= 145  Num columns= 289    received:  2003177 155852Z
v-component of wind
GRIB ID numbers: Geographic =3744; PAR = 34; Model ID = 96; Level type =100
Units of gridded variable are MPS  Scale of variable is:   2
Mercator Projection
Min Lat= -90.00  Max Lat=  90.00  Min Lon=-180.00  Max Lon= 180.00
Latitude Increment= 1.2500  Longitude Increment= 1.2500
Number of grids listed = 2
GRDLIST - done


***** Notice *****

- the 'Geographic =' for the grids copied from adde.ucar.edu are '43'
  for the first grid, and '44' for the second grid

- the 'Geographic =' for the grids copied from the SSEC server are
  both '3744'

Here is what is going on.  The AVN grids are pieced together from 8
octet grids each of which as a separate geographic ID: 37, 38, ...,
44.  The geographic ID that gets assigned to the stitched together grid
that the v2002 XCD grib decoder creates is either the first or last (I
don't remember which) octet grid that was stitched.  Since the order
that the octet grids are received in NOAAPORT is not fixed, one can not
be sure of which geographic ID will be assigned to the stitched
together composite grid.  This accounts for the DERIVE=ABV working
sometimes and not others!

SSEC made a change in the v2003 version of XCD to set the geographic ID
to a fixed number '3744' (representing any/all of 37, ..., 44).

The selection code in McIDAS goes to great pains to make sure that the
grids being operated on are exactly the same.  It is exiting out the
ABV calculation as soon as it sees that the geographic ID for the two
grids does not match.

I verified that if I change the geographic ID in the first grid copied
from adde.ucar.edu (U, which was '43') to match that in the second grid
copied from adde.ucar.edu (V, which is '44') then the DERIVE=ABV
function works as it designed to.

I see no _easy_ way to get around this problem from the user side.
You could copy over the grids you want to use from the server one
at a time and change the value in the grid so that hte selection will
match, but that is kinda ugly.  Even though I was not quite ready to
load the v2003 version of XCD on adde.ucar.edu, I may well do so
to help you get around this problem.

More later...

Tom