20020502: Unidata gribdec.tar.Z bundle (cont.)

>From: "Hoeth, Brian" <address@hidden>
>Organization: JSFC
>Keywords: 200205011845.g41IjMa25331 McIDAS gribdec.tar.Z

Brian, et. al.,

In a previous Brian noted:

>(2) The decoder doesn't seem to decode parameters of type "surface" that
>have a "timerange".  For example, I went and got the 00 hour forecast data
>for the 12Z run of the 8km data mentioned below.  The GRIB file
>(AWPREG00.tm00) has 143 records in it, but only 135 are decoded into GRIDS.
>When I examined which parameters were missed, I found that 7 of the 8
>missing parameters were "surface" parameters with a "timerange":
>
>(Note: timerange values in parentheses are what is gotten from running wgrib
>-V on the GRIB file)
>
>(record 14) APCP sfc 0-0hr (timerange 4)
>(record 15) ACPCP sfc 0-0hr (timerange 4)
>(record 16) WEASD sfc 0-0hr (timerange 4)
>(record 119) NCPCP sfc 0-0hr (timerange 4)
>(record 120) SNOM sfc 0-0hr (timerange 4)
>(record 124) LHTFL sfc 0-0hr (timerange 3)
>(record 125) SHTFL sfc 0-0hr (timerange 3)

I grabbed the file:

ftp://ftpprd.ncep.noaa.gov/pub/emc/mmb/mmbpll/east08.t18z/AWPREG00.tm00

and ran gribdec.k through its paces.  I found that the 10 out of 143
gribs were not decoded.  Its pretty clear that the file I grabbed was
not the same as the one you grabbed, so my findings may not match yours.

I found that 9 of the gribs that weren't decoded had constant values,
and Mcgribdecoder.c was not handling the case where grids could be
constant fields.  FYI, the gribs with constant data were all
precipitation grids of one sort or another in the file I grabbed.

I updated Mcgribdecoder.c to correctly decode constant fields.

>(3) The decoder also doesn't seem to decode parameters which are at a
>specific level below ground.  As mentioned above, there were 8 parameters
>that weren't decoded and 7 are listed above, the last parameter was:
>
>(record 143) TSOIL 300 cm down 

The 10th grid that I could not decod was the same as reported here: STEM
(soil temperature).  The problem here was that Mcmkmcgrid.c did not
have a clause that could handle a vertical level type of 111.  Since I
was already mucking around with Mcmkmcgrid.c, I added support for the
following levels:

level code  level                           unit
-----------+-------------------------------+-------------------------------
  111       depth below land surface        cm
  113       isentropic (theta) level        K
  125       specified height above ground   cm
  160       depth below sea level           m

There are still a number of other levels listed in GRIB Edition 1 (FM92)
that could be defined, but these will have to wait for another day.

I stuck a new gribdec.tar.Z file out in the unix/gribdec directory of
our passworded McIDAS FTP account on our FTP server, ftp.unidata.ucar.edu
for you to grab.

Later...

Tom