[Met_help] ASCII point observation format
John Halley Gotway
johnhg at rap.ucar.edu
Mon Dec 15 12:13:51 MST 2008
The table of message types can be found here in the PREPBUFR documentation: http://www.emc.ncep.noaa.gov/mmb/data_processing/prepbufr.doc/table_1.htm
For the "surface" observations - meaning 2-meter variables (like temp, dpt, pressure, rh) and 10-meter winds, use a message type of "ADPSFC". MET matches point observations to the surface in a very
simplistic way... we're just mimicking the way NCEP does it. Basically, if an observation has a message type of ADPSFC or SFCSHP, it is considered to be at the surface. So when verifying 2-meter
temperature or 10-meter winds, the Point-Stat tool in MET only uses observations with one of those message types.
Yes, just use the code from the GRIB file output of WPP. For your reference, here's a table of GRIB codes: http://www.nco.ncep.noaa.gov/pmb/docs/on388/table2.html
Also, the "wgrib" command is very help for dumping the contents of a GRIB file to the screen. It will tell you what GRIB codes are contained in a file.
Since you're just verifying surface variables, I believe that you could put in a value of -9999 for missing data.
Again, just put in a value of -9999 for missing data. MET does not currently use this column of data anyway. Since we're just doing the simplistic thing of matching to surface observations based on
the message type, we're not using the "Height" information for anything. We put that column in as a placeholder to allow us to do more sophisticated matching by vertical level in the future.
Listed below is a sample ASCII observation file I use for testing. The numbers don't mean much since it's just for software testing. But the sample observations are for 12-hour accumulated
precipitation at the surface:
ADPSFC SID1 20050807_120001 40.00 -105.00 2001 61 12 -9999 1.01
ADPSFC SID2 20050807_120002 40.00 -105.00 2002 61 12 -9999 1.02
ADPSFC SID3 20050807_120003 40.00 -105.00 2003 61 12 -9999 1.03
ADPSFC SID4 20050807_120004 40.00 -105.00 2004 61 12 -9999 1.04
ADPSFC SID5 20050807_120005 35.00 -105.00 2005 61 12 -9999 1.05
ADPSFC SID6 20050807_120006 35.00 -105.00 2006 61 12 -9999 1.06
ADPSFC SID7 20050807_120007 35.00 -105.00 2007 61 12 -9999 1.07
ADPSFC SID8 20050807_120008 35.00 -105.00 2008 61 12 -9999 1.08
ADPSFC SID9 20050807_120009 35.00 -105.00 2009 61 12 -9999 1.09
Hope that helps,
John Halley Gotway
Mark Seefeldt wrote:
> I am attempting to create my own ASCII files to use with ascii2nc to
> create an observation dataset to use with MET. I am working with two
> sets of data. One is surface observations at two locations with the
> basic state variables (t2m, td2m, pres, rh, u, v) and also radiation
> variables (SW_d, LW_d). The second set of data is upper air
> observations. The sample included in the METv1.1 distribution appears
> to be for an upper air dataset, so I can likely figure that out in the
> future. In the current, I am having some confusion with a few of the
> columns in the ASCII Point Observation Format:
> column 1 - message_type - How do I know what I am supposed to fill in
> for the message_type? Is there a table somewhere with the available
> message_types listed?
> column 7 - I am creating the GRIB files using WPP so I will just enter
> the corresponding values for the observations into this column.
> column 8 - The MET User's Guide says: "Pressure level in hPa or
> accumulation interval in hours for the observation value." What do I
> fill in for the T_2m, Td_2m, u_10m, v_10m, or SW_d?
> column 9 - "Height in msl of the observation value."
> -Similar to column 8, what do I fill in for the T_2m, Td_2m, u_10m,
> v_10m, or SW_d?
> If you can provide me a sample of an ASCII file for surface observations
> it will likely help me.
> Met_help mailing list
> Met_help at mailman.ucar.edu
More information about the Met_help