[Met_help] min/max verification

John Halley Gotway johnhg at ucar.edu
Fri Sep 4 08:28:42 MDT 2009


Holly,

Thanks for sending the sample data.  The reason you're not getting any matches is that the times aren't matching up.  Based on the wgrib output, it looks like the forecast you're verifying is
initialized at 20090901 00Z and valid between 20090901 12Z and 20090902 00Z.

Unfortunately, the Point-Stat tool isn't handling the times the way you'd expect.  Instead, it's interpreting the "valid time" of the forecast to be 20090902 00Z and using the beg_ds and end_ds
configuration parameters (set to -900 and 900 seconds in the file you sent).  So Point-Stat is only looking in the time window of 20090901 23:45 to 20090902 00:15 for matches, but it doesn't find any.

So we're not doing a good job handling forecasts over a time range.  To my knowledge, we haven't had users verifying these types of forecasts in the past.  Instead, they've all been instantaneous
forecasts.  So we'll need to figure out how to address this time matching.

In the meantime, please try editing you Point-Stat configuration file as follows:
beg_ds = -43200
end_ds = 0

Since Point-Stat thinks the "valid" time is 20090902 00Z, these settings will define a matching time window of 20090901 12Z to 20090902 00Z.

Thanks,
John

Holly Hassenzahl wrote:
> Hi John-
> 
>  
> 
> I hope this email finds you well!  We've been running into some issues
> with trying to verify the min and max temps.  I have uploaded 3 files
> into my directory on the ftp site
> (incoming/irap/met_help/hassenzahl_data):
> 
>  
> 
> fcst00z_12L_min.grib  (NDFD forecast file)
> 
> obs_17z.nc  (ASCII to NetCDF output of the obs...this file holds the max
> and min from yesterday, as well as the 17z temp)
> 
> PointStatConfig_ndfd_min
> 
>  
> 
> For some reason, MET can't find any pairs.  I'm wondering if it has
> something to do with the way NDFD specifies the valid forecast times?
> For example, they say that TMIN is valid for the 12-24hr forecast and
> TMAX is valid for the 24-36hr forecast.  Instead of at 12hr and 24hr,
> respectively.
> 
>  
> 
> Would you mind taking a look at this for me and see if anything jumps
> out at you?  My sincere thanks,
> 
> Holly
> 
>  
> 
> ___________________________________
> 
> Holly C. Hassenzahl
> 
> Meteorologist, Science Analyst
> 
> Data Products Group
> 
>  
> 
> Weather Central, Inc.
> 
> 401 Charmany Drive Suite 200
> 
> Madison, WI 53711
> 
> +1.608.274.5789
> 
> +1.608.276.4600
> 
> http://www.weathercentral.tv
> 
>  
> 
>  
> 
> 


More information about the Met_help mailing list