[Met_help] [rt.rap.ucar.edu #69601] History for pointwise wind direction

[John Halley Gotway via RT]

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

Hello,

I wanted to ask whether there is a possibility to verify wind direction
pointwise, so eg. calculate RMSE of wind direction for every single
observation point, and then make and average of these RMSE values.

If I make the verification following the Users Manual (point_stat and then
stat_analysis) then it calculates the error of area averaged forecast and
observed wind direction (if I understand correctly). I guess this method
favours low resolution models (I am comparing low and high resolution
models).

Thanks and best regards,
Balazs Szintai


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

Subject: pointwise wind direction
From: John Halley Gotway
Time: Tue Nov 04 11:00:01 2014

Balazs,

We added some functionality in MET version 5.0 to make this easier.
Each
time you run point-stat, verify the U and V components (UGRD and VGRD)
and
dump out the matched pairs (MPR) line type.

Then run the following type of stat-analysis job:
  stat_analysis -lookin path/to/stat/data -job aggregate_stat
-line_type
MPR -out_line_type WDIR -by OBS_SID

That job reads through the input STAT files, finds pairs of UGRD/VGRD
lines, sorts them by the observation station id column (OBS_SID), and
the
computes wind direction errors for each station.  If you data contains
UGRD/VGRD pairs for different vertical levels, lead times, or masking
regions, you can specify which ones to use on the command line.

Hope that helps.

Thanks,
John Halley Gotway
met_help at ucar.edu

On Tue, Nov 4, 2014 at 9:27 AM, Balazs Szintai via RT
<met_help at ucar.edu>
wrote:

>
> Tue Nov 04 07:27:39 2014: Request 69601 was acted upon.
> Transaction: Ticket created by balazs.szintai at gmail.com
>        Queue: met_help
>      Subject: pointwise wind direction
>        Owner: Nobody
>   Requestors: balazs.szintai at gmail.com
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69601 >
>
>
> Hello,
>
> I wanted to ask whether there is a possibility to verify wind
direction
> pointwise, so eg. calculate RMSE of wind direction for every single
> observation point, and then make and average of these RMSE values.
>
> If I make the verification following the Users Manual (point_stat
and then
> stat_analysis) then it calculates the error of area averaged
forecast and
> observed wind direction (if I understand correctly). I guess this
method
> favours low resolution models (I am comparing low and high
resolution
> models).
>
> Thanks and best regards,
> Balazs Szintai
>
>

------------------------------------------------
Subject: pointwise wind direction
From: Balazs Szintai
Time: Fri Nov 07 07:15:25 2014

Thanks a lot, John. It is working now.

Balazs




On Tue, Nov 4, 2014 at 7:00 PM, John Halley Gotway via RT
<met_help at ucar.edu
> wrote:

> Balazs,
>
> We added some functionality in MET version 5.0 to make this easier.
Each
> time you run point-stat, verify the U and V components (UGRD and
VGRD) and
> dump out the matched pairs (MPR) line type.
>
> Then run the following type of stat-analysis job:
>   stat_analysis -lookin path/to/stat/data -job aggregate_stat
-line_type
> MPR -out_line_type WDIR -by OBS_SID
>
> That job reads through the input STAT files, finds pairs of
UGRD/VGRD
> lines, sorts them by the observation station id column (OBS_SID),
and the
> computes wind direction errors for each station.  If you data
contains
> UGRD/VGRD pairs for different vertical levels, lead times, or
masking
> regions, you can specify which ones to use on the command line.
>
> Hope that helps.
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
> On Tue, Nov 4, 2014 at 9:27 AM, Balazs Szintai via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > Tue Nov 04 07:27:39 2014: Request 69601 was acted upon.
> > Transaction: Ticket created by balazs.szintai at gmail.com
> >        Queue: met_help
> >      Subject: pointwise wind direction
> >        Owner: Nobody
> >   Requestors: balazs.szintai at gmail.com
> >       Status: new
> >  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=69601 >
> >
> >
> > Hello,
> >
> > I wanted to ask whether there is a possibility to verify wind
direction
> > pointwise, so eg. calculate RMSE of wind direction for every
single
> > observation point, and then make and average of these RMSE values.
> >
> > If I make the verification following the Users Manual (point_stat
and
> then
> > stat_analysis) then it calculates the error of area averaged
forecast and
> > observed wind direction (if I understand correctly). I guess this
method
> > favours low resolution models (I am comparing low and high
resolution
> > models).
> >
> > Thanks and best regards,
> > Balazs Szintai
> >
> >
>
>

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