20000707: SAUS45 Reports (fwd)

---------- Forwarded message ----------
Date: Fri, 07 Jul 2000 11:05:54 -0600
From: Peter Neilley <address@hidden>
Subject: SAUS45 Reports

Larry's message about the Wyoming DOT sites reminded me about something I've 
been
meaning to post here for awhile regarding METAR's in the SAUS45 and SXUS97 
collectives.
Care and caution needs to be used when dealing with these "extra" METARs.  I 
think
some of this info was posted in a thread awhile back, so sorry if this is not
all new info.  Note that Larry referred to obs under the SDUS45 headings.... 
many
(if not all) of those obs are also carried under the SAUS45 heading, so the info
posted here also pertains to Cheyenne's DOT observations.

These bulletins seem to contain METAR's broadcast from local NWS offices, mostly
in the western US.  In many cases, the METARs contain reports that are not 
broadcast
in the normal (e.g. SAUS80) collectives.  Thus if one is looking for the 
greatest
possible observational coverage, grabing and using these "extra" METARs can 
make a
big difference in data sparce regions.  However, these reports have some quirks 
about 
them which can cause difficulties.

First of all, station IDs are not unique.  For example, the Rapid City office
issues a METAR for station KSLK, but this ID is really for Saranac Lake NY.   
For
example, from 10Z today, KUNR (Rapid City NWS) issued:

  KSLK 070947Z AUTO 00000KT 17/11 RMK A01 Pk WND MM=

while the regular KSLK metar (in the SAUS70 collective) was

  KSLK 070951Z AUTO VRB03KT 5SM BR FEW005 OVC037 06/06 A3001 RMK AO2 SLK170 
T006100056=

Thus if you combine these data into one database (e.g. GEMPAK file) and plot the
data, confusing results often occur if the ob from the Rapid City office gets 
plotted
in upstate NY.  Other known duplicate IDs are KMTR (from Cheyenne duplicate with
Monterey CA) and KPHP (from Goodland, duplicate with Philip SD).

Also, the times in the WMO headers for these bulletins often does not reflect 
the observation
hour.  The hour in the WMO header is not "rounded off" to the nearest hour like 
most
METAR collectives.    So if a ob is taken at 0957Z, it is sent with a WMO 
header indicating
9Z, not 10Z like most of the normal (e.g. SAUS80) METARs.  Thus if you use just 
the hour
in the WMO header as the way in which data is sorted into times, you can often 
get a
mixed bag of obs from the same site.  If you then plot the obs from one hour, 
the data
may not all be syncronized.

For example, obs from Limon Colorado with 10Z headers today were broadcast 
under each of the
following headers:

  SAUS80 KWBC 071000
     KLIC 070955Z AUTO 02004KT 10SM CLR 06/04 A3013 RMK A02 SLP138 T00560044 
TSNO

  SAUS45 KPUB 071057
     KLIC 071055Z AUTO 36003KT 10SM CLR 06/05 A3014 RMK A02 SLP140 T00610050 
TSNO

  SAUS45 KBOU 071057 COR
  MTRLIC
     KLIC 071055Z AUTO 36003KT 10SM CLR 06/05 A3014 RMK A02 SLP140 T00610050 
TSNO

Clearly the latter two obs were really 11Z obs (and indeed were contained in the
SAUS80 071100 collective).

But, if you stuffed all of the obs that had 10Z in the header into one 10Z 
file, you would
have obs for both (nominally) 10Z and 11Z obs if you had the SAUS45 collectives.


Peter Neilley, NCAR/RAP
address@hidden
303-497-8446