20010329: Decoding special observations via NOAAPORT data stream

Scott,

The default maximum number of times allowed in a GEMPAK file is 300
as set in $GEMPAK/include/GEMPRM.PRM and gemprm.h. Your use of
-m 1440 is greater than this limit. Exceeding the limit 
may cause all sorts of problems with sizes.

Also, by using -m 72 in my example, the decoder will store
data using 20 minute bins instead of rounding to the hour.
This is useful for some locations which take observations 3x per hour
such as the Minnesota and Iowa mesonets, AWOS and other reporting sites.
72 is the only number recognized by dcmetr for producing 20 minute bins
(eg 24 hours * 3 per hour). Any other number will be ignored in
terms of 20 minute bins (dcmetr -help).

Up to 3 specials will be stored in the gempak file for any hour per station
and you can view the raw reports with PARM=TEXT;SPCL in sflist. The specials 
won't be decoded.

Steve Chiswell
Unidata User Support



>From: Scott Halvorson <address@hidden>
>Organization: UCAR/Unidata
>Keywords: 200103291650.f2TGo8L14914

>Hi,
>
>I am having trouble decoding special observations (hourly observations
>work) from a NOAAPORT data stream using LDM and the Gempak decoder
>"dcmetr".  The surface obs decoding  section of our pqact.conf file is:
>
># Section 4.  Realtime decoding into GEMPAK files.
>#
># The Domestic Surface Hourly Observations
>#
>WMO     ^S[AP].* .... ([0-3][0-9])([0-2][0-9])([0-5][0-9])  PIPE
>        /usr/local/ldm/decoders/dcmetr -b 9  -m 1440 -s sfmetar_sa.tbl
>        -d /usr/local/ldm/logs/dcmetr.log
>        -e GEMTBL=/home/ldm/gempak-5.6/gempak/tables
>        /ext_obs/gempak/data/YYYYMMDD.metar.gem
>#
>
> An example of the errors showing up in the log file are:
>
>   [1384] 010328/1429 [DCMETR 3] SAUS70  KWBC    282117
>   [1384] 010328/1429 [DCMETR 4]
>   [1384] 010328/1429 [DC -16]
>
>Where,
>
>DCMETR
>   +4      ! Invalid date/time:  !AS
>   +3      ! Invalid bulletin:  !AS
>DC
>    -16     ! Invalid time
>
>Here is your example of a pqact.conf file that excepts both hourly and
>special observations.  The only difference that I can see is this is for
>DDS instead of WMO.  Note also that ([0-5][0-9]) was appended in our
>pqact.conf above, but with or without we still had the same problem.
>
>#
># sfc obs and specials
># Use -m 72 to store data in 20 minute bins, rather than hourly.
># 20 minute bins useful for mesonet and AWOS data
>#
>DDS|IDS ^S[AP].* .... ([0-3][0-9])([0-2][0-9])
>        PIPE    decoders/dcmetr -b 9 -m 72 -s sfmetar_sa.tbl
>        -d data/gempak/logs/dcmetr.log
>        -e GEMTBL=/home/gempak/NAWIPS/gempak/tables
>        data/gempak/surface/YYYYMMDD_sao.gem
>
>
>Hourly observations are decoding okay in the Gempak file, but the time
>stamp appears to be rounded to the nearest hour.
>
>Thanks for you time and any help.
>
>Scott Halvorson
>
>
>
>