[Met_help] [rt.rap.ucar.edu #80021] History for threshold for MODE tool

[John Halley Gotway via RT]

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

Hello, I have a question about setting the threshold in MODE config file.
The verification is for wind speed at 10m, I only want to compare the
strong wind(>=12m/s) features between obs and fcst, I feel confuse about
the raw_thresh and conv_thresh. I understand the conv_radius and
conv_threshold are the two parameters controlling MODE. But should I set
raw_thresh to >=12.0 as well? I tried to set raw_threshold to different
value and ran MODE tool, such as >=5.0, >=8.0, the results have slight
differences.
*****************************************
 raw_thresh        = >=0.0;
   conv_thresh       = >=12.0;
*******************************************

Thanks for your help!

-- 
Ying Song, PhD, Post Doctoral Fellow
Department of Earth and Atmospheric Sciences
Saint Louis University

ysong4 at slu.edu


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

Subject: threshold for MODE tool
From: John Halley Gotway
Time: Thu Mar 30 19:18:51 2017

Ying,

No, you can ignore the raw_thresh and just use the conv_thresh to
define
the objects.

The raw_thresh is very seldom needed.  After reading the raw data
values,
the raw_thresh is applied and any values not meeting that threshold
criteria are reset to a value of 0.

Why would you ever need this functionality?  In the past, we applied
MODE
to forecasts of reflectivity.  But the forecasts were only for
reflectivity
>= 35 dBZ.  So the forecast contained areas of reflecitivity >= 35
surrounded by values of 0.  The observations contained the full range
of
observed reflectivity values... so the 35 dBZ observations were
surround by
reflectivity less than 35.

During the convolution smoothing process, MODE averages the value at
each
grid point with values nearby.  In the forecast, the nearby values
were 0
while in the obs, the nearby values were in the 30's.  Ultimately,
this
yielded much smaller forecast objects in MODE (because of all those
0's).
We used the raw_thresh to zero out any observation values < 35 so that
the
convolution smoother step would be applied to the forecast and
observation
data fairly.

So as you can see, this is pretty specialized and seldom used.

Hope that helps clarify.

Thanks,
John

On Thu, Mar 30, 2017 at 6:34 PM, Ying Song via RT <met_help at ucar.edu>
wrote:

>
> Thu Mar 30 18:34:27 2017: Request 80021 was acted upon.
> Transaction: Ticket created by ysong4 at slu.edu
>        Queue: met_help
>      Subject: threshold for MODE tool
>        Owner: Nobody
>   Requestors: ysong4 at slu.edu
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=80021 >
>
>
> Hello, I have a question about setting the threshold in MODE config
file.
> The verification is for wind speed at 10m, I only want to compare
the
> strong wind(>=12m/s) features between obs and fcst, I feel confuse
about
> the raw_thresh and conv_thresh. I understand the conv_radius and
> conv_threshold are the two parameters controlling MODE. But should I
set
> raw_thresh to >=12.0 as well? I tried to set raw_threshold to
different
> value and ran MODE tool, such as >=5.0, >=8.0, the results have
slight
> differences.
> *****************************************
>  raw_thresh        = >=0.0;
>    conv_thresh       = >=12.0;
> *******************************************
>
> Thanks for your help!
>
> --
> Ying Song, PhD, Post Doctoral Fellow
> Department of Earth and Atmospheric Sciences
> Saint Louis University
>
> ysong4 at slu.edu
>
>

------------------------------------------------
Subject: threshold for MODE tool
From: Ying Song
Time: Fri Mar 31 16:45:18 2017

okay. got it. Thanks for the reply! I will just ignore the raw_thresh
parameter.

On Thu, Mar 30, 2017 at 8:18 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:

> Ying,
>
> No, you can ignore the raw_thresh and just use the conv_thresh to
define
> the objects.
>
> The raw_thresh is very seldom needed.  After reading the raw data
values,
> the raw_thresh is applied and any values not meeting that threshold
> criteria are reset to a value of 0.
>
> Why would you ever need this functionality?  In the past, we applied
MODE
> to forecasts of reflectivity.  But the forecasts were only for
reflectivity
> >= 35 dBZ.  So the forecast contained areas of reflecitivity >= 35
> surrounded by values of 0.  The observations contained the full
range of
> observed reflectivity values... so the 35 dBZ observations were
surround by
> reflectivity less than 35.
>
> During the convolution smoothing process, MODE averages the value at
each
> grid point with values nearby.  In the forecast, the nearby values
were 0
> while in the obs, the nearby values were in the 30's.  Ultimately,
this
> yielded much smaller forecast objects in MODE (because of all those
0's).
> We used the raw_thresh to zero out any observation values < 35 so
that the
> convolution smoother step would be applied to the forecast and
observation
> data fairly.
>
> So as you can see, this is pretty specialized and seldom used.
>
> Hope that helps clarify.
>
> Thanks,
> John
>
> On Thu, Mar 30, 2017 at 6:34 PM, Ying Song via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > Thu Mar 30 18:34:27 2017: Request 80021 was acted upon.
> > Transaction: Ticket created by ysong4 at slu.edu
> >        Queue: met_help
> >      Subject: threshold for MODE tool
> >        Owner: Nobody
> >   Requestors: ysong4 at slu.edu
> >       Status: new
> >  Ticket <URL:https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__rt.rap.ucar.edu_rt_Ticket_Display.html-3Fid-3D80021&d=
> DwIDaQ&c=Pk_HpaIpE_jAoEC9PLIWoQ&r=DoGUCazxekmrzloEJ54sJA&m=
>
L1xwn2YVHPwc5amZwUVHSh31fHSjPWhhTHzYNwSVGE0&s=vgmEHnRo17JIEZg106IyXTTvJG-
> 4w43cfiAC5KR1NQE&e= >
> >
> >
> > Hello, I have a question about setting the threshold in MODE
config file.
> > The verification is for wind speed at 10m, I only want to compare
the
> > strong wind(>=12m/s) features between obs and fcst, I feel confuse
about
> > the raw_thresh and conv_thresh. I understand the conv_radius and
> > conv_threshold are the two parameters controlling MODE. But should
I set
> > raw_thresh to >=12.0 as well? I tried to set raw_threshold to
different
> > value and ran MODE tool, such as >=5.0, >=8.0, the results have
slight
> > differences.
> > *****************************************
> >  raw_thresh        = >=0.0;
> >    conv_thresh       = >=12.0;
> > *******************************************
> >
> > Thanks for your help!
> >
> > --
> > Ying Song, PhD, Post Doctoral Fellow
> > Department of Earth and Atmospheric Sciences
> > Saint Louis University
> >
> > ysong4 at slu.edu
> >
> >
>
>


--
Ying Song, PhD, Post Doctoral Fellow
Department of Earth and Atmospheric Sciences
Saint Louis University

ysong4 at slu.edu

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