[Met_help] [rt.rap.ucar.edu #81276] History for Climodata to be used in Point Stat

[John Halley Gotway via RT]

----------------------------------------------------------------
  Initial Request
----------------------------------------------------------------

Hi John, I am verifying precipitation probabilities using Point Stat against
observations and want to use climo to get the Brier Skill Scores.  The climo
file I am using is the observational frequency of the event that I write out
to a NETCDF file using ascii2nc so I have a file of observations where they
all have the same observational frequency of the event.  The first question
since the user manual is vague on this:

1)  Does the climo file have to match the observation file as are as
dimensions and grib codes (that is how I am currently doing) or does the
climo file need to be a 2D array of the observation frequency at each grid
point matching the fcst grid?

If the answer above is it as to match the observation file, then what field
parameters do I use?  The manual says to match of obs field parameters
(which is what I am doing) but I always get the message "found 0 climatology
fields".  

2) If climo has to match the fcst grid, then I assume I would have to write
out a general NETcdf file of a grid with the one obs freq value at every
grid point.  The variable name in the NETcdf file would have to match the
fcst grid ("APCP" in this case).  Then the field dictionary for climo would
match the fcst field dictionary.  Is this correct?

Attached is the config file.  The command line I am running is:

/h/WXQC/met-6.0/bin/point_stat
/h/data/global/WXQC/data/geps_verify/qcsave/2017070900/grib2.2017070900.0012
.NOAA /h/data/global/WXQC/data/met/nc_obs/precip/nc_2017070912
/h/WXQC/met-6.0/data/config/PointStatConfig_geps_precip_updated -outdir
/h/data/global/WXQC/data/met/ptstat/mdlob_pairs -v 4

I placed the grid, obs, and climo file on the server.

Your help would be greatly appreciated.

Thanks
Bob Craig


----------------------------------------------------------------
  Complete Ticket History
----------------------------------------------------------------

Subject: Climodata to be used in Point Stat
From: John Halley Gotway
Time: Fri Jul 21 15:15:07 2017

Bob,

I see you have some questions about using climatology data in Point-
Stat.

MET expects and processes climatology data as a gridded data file, not
as
point observations.  I see that you have the climo data in
3046a495-e34c-4f93-8b57-489eb58388b9 formatted in the point
observation
format.  That will not work with the current logic.  It would need to
be a
gridded dataset instead.

Looking at your config file, I had 2 suggestions:

In met-6.0, instead of:
   cat_thresh =
[>=0,>=0.1,>=0.2,>=0.3,>=0.4,>=0.5,>=0.6,>=0.7,>=0.8,>=0.9,>=1.0];
You can simple specify:
   cat_thresh = [ ==0.1 ];

Either way works, but the second is shorter.

You have a typo in the mask section on the side line:
   sid     = "[]";
Should be:
   sid     = [];

Otherwise, it's looking a station id named as "[]" and that'll result
in 0
matched pairs.

Thanks,
John




On Wed, Jul 19, 2017 at 3:09 PM, robert.craig.2 at us.af.mil via RT <
met_help at ucar.edu> wrote:

>
> Wed Jul 19 15:09:45 2017: Request 81276 was acted upon.
> Transaction: Ticket created by robert.craig.2 at us.af.mil
>        Queue: met_help
>      Subject: Climodata to be used in Point Stat
>        Owner: Nobody
>   Requestors: robert.craig.2 at us.af.mil
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=81276 >
>
>
> Hi John, I am verifying precipitation probabilities using Point Stat
> against
> observations and want to use climo to get the Brier Skill Scores.
The
> climo
> file I am using is the observational frequency of the event that I
write
> out
> to a NETCDF file using ascii2nc so I have a file of observations
where they
> all have the same observational frequency of the event.  The first
question
> since the user manual is vague on this:
>
> 1)  Does the climo file have to match the observation file as are as
> dimensions and grib codes (that is how I am currently doing) or does
the
> climo file need to be a 2D array of the observation frequency at
each grid
> point matching the fcst grid?
>
> If the answer above is it as to match the observation file, then
what field
> parameters do I use?  The manual says to match of obs field
parameters
> (which is what I am doing) but I always get the message "found 0
> climatology
> fields".
>
> 2) If climo has to match the fcst grid, then I assume I would have
to write
> out a general NETcdf file of a grid with the one obs freq value at
every
> grid point.  The variable name in the NETcdf file would have to
match the
> fcst grid ("APCP" in this case).  Then the field dictionary for
climo would
> match the fcst field dictionary.  Is this correct?
>
> Attached is the config file.  The command line I am running is:
>
> /h/WXQC/met-6.0/bin/point_stat
>
/h/data/global/WXQC/data/geps_verify/qcsave/2017070900/grib2.2017070900
> .0012
> .NOAA /h/data/global/WXQC/data/met/nc_obs/precip/nc_2017070912
> /h/WXQC/met-6.0/data/config/PointStatConfig_geps_precip_updated
-outdir
> /h/data/global/WXQC/data/met/ptstat/mdlob_pairs -v 4
>
> I placed the grid, obs, and climo file on the server.
>
> Your help would be greatly appreciated.
>
> Thanks
> Bob Craig
>
>

------------------------------------------------