[Met_help] PBN2NC chocking

John Halley Gotway johnhg at ucar.edu
Tue Oct 27 08:35:49 MDT 2009


Graziano,

Hmmm.  Which files are you using from that site?

I just retrieved the following file:
http://nomads.ncdc.noaa.gov/data/gdas/200910/20091024/gdas1.t00z.prepbufr.nr

And ran it through PB2NC.  The good news is that it ran fine without error.  The bad news is that it took 6 minutes and 44 seconds to run!  We realize that PB2NC takes a long time to run, and we're
hoping to find ways of speeding it up.  However, there really are a lot of observations to process.  This file I ran contains 6 hours worth of global observations.  PB2NC had to parse through
1,829,960 observation values.  So that's what's taking so long!

You should be looking for PrepBufr files following this naming convention:
http://nomads.ncdc.noaa.gov/data/gdas/YYYYMM/YYYYMMDD/gdas1.tHHz.prepbufr.nr, where HH is 00, 06, 12, or 18.  The observations are centered on this time.  So each GDAS file contains observations +/- 3
hours around the time listed.  For example, the 12Z file contains observations between 09Z and 15Z.

Please note that the files named "gdas1.tHHz.prepbufr.unblok.nr" should work as well.  They should be identical to the files without the "unblok" in them except for a few bytes in the beginning.  The
observations in them should be identical.  And PB2NC should be able to read both files fine.

Please let me know if you continue to experience problems.

Thanks,
John Halley Gotway
johnhg at ucar.edu

Graziano Giuliani wrote:
> Hi to all,
> first time MET (2.0, latest patches applied) user seeking for help. I have
> downloaded prepbufr data from:
> 
> http://nomads.ncdc.noaa.gov/data/gdas/200910/20091024
> 
> either upper air or surface have same problem:
> 
> ERROR: process_pbfile() -> the observation time should remain the same for all
> PrepBufr messages: 1256338800 != 1256342400
> 
> I checked that:
> 
> date --date "1970-01-01 00:00:00 UTC + 1256338800 seconds"
> Fri Oct 23 23:00:00 UTC 2009
> 
> date --date "1970-01-01 00:00:00 UTC + 1256342400 seconds"
> Sat Oct 24 00:00:00 UTC 2009
> 
> So I have a problem, as it seems not true that the observation time remains the
> same for all PrepBufr messages in a file.
> 
> How to solve this?
> 
> Thanks,
> 
> Graziano.
> _______________________________________________
> Met_help mailing list
> Met_help at mailman.ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/met_help


More information about the Met_help mailing list