[Met_help] [rt.rap.ucar.edu #75882] History for hello, sir. qustion about threshold in modeconfig file
John Halley Gotway via RT
met_help at ucar.edu
Tue Apr 26 14:00:08 MDT 2016
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hello,sir:
In MODEConfig file I want set raw threshld as "75<=thre <100", how would I set in MODEConfig file. I've tried .And there came out a syntax error.
Thank u very much and i'm looking forward to your early reply.
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: hello,sir. qustion about threshold in modeconfig file
From: John Halley Gotway
Time: Tue Apr 12 10:00:19 2016
Hello, I see you have question about specifying a combination of
thresholds. The correct syntax for this is shown below:
raw_thresh = >=75&&<100;
That selects all values between 75 and 100.
This is mentioned on page 44 of the MET users's guide:
http://www.dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v5.1.pdf
Lastly, I'm wondering why you're choosing a raw threshold for MODE at
all.
Typically users do not need to specify a raw_thresh value. All it
does is
set any values which do not meet that threshold criteria to a value of
0.
For example, we ran MODE to compare a forecast of reflectivity to
observed
reflectivity. However, the forecasts were only for reflectivity
values of
35 dBZ and higher while the observed reflectivity contained a full
range of
values. Using MODE to compare forecast areas of reflectivity 35 dBZ
and
higher to a full range of reflectivity values isn't really "fair".
The
convolution, or smoothing, step will lead to smaller forecast objects
since
the 35+ values are being smoothed with surrounding values of 0.
Therefore,
we set the observed raw_thresh to >=35. That made it so that both the
forecast and observation fields had values of 35 dBz surrounded by
values
of 0.
Unless you're in a similar situation, where the range of values in the
forecast and observation field differ, I doubt the raw_thresh is the
correct thing to be using.
Hope that helps clarify.
Thanks,
John Halley Gotway
On Tue, Apr 12, 2016 at 6:20 AM, Zhuang from Nuist via RT
<met_help at ucar.edu
> wrote:
>
> Tue Apr 12 06:20:55 2016: Request 75882 was acted upon.
> Transaction: Ticket created by zxrxz3212009 at 163.com
> Queue: met_help
> Subject: hello,sir. qustion about threshold in modeconfig file
> Owner: Nobody
> Requestors: zxrxz3212009 at 163.com
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
>
>
> Hello,sir:
> In MODEConfig file I want set raw threshld as "75<=thre
<100",
> how would I set in MODEConfig file. I've tried .And there came out
a
> syntax error.
> Thank u very much and i'm looking forward to your early
reply.
>
------------------------------------------------
Subject: hello,sir. qustion about threshold in modeconfig file
From: Zhuang from Nuist
Time: Tue Apr 12 19:46:24 2016
Dir sir:
Thanks very much for your kind consideration.
As your said, compare objects with a fixed threshold is not
"fair", that's to say, if I want to compare the mesoscale system in a
rain case(for eg. >=50mm) is not resonable. But references like
Ebert(2008); William(2009) all chosed a fixed threshold in order to
compare the object, like CRA method.
Also, can mode use a percentile threshold instead of a fixed
one?
Thanks very much!
At 2016-04-13 00:00:19, "John Halley Gotway via RT"
<met_help at ucar.edu> wrote:
>Hello, I see you have question about specifying a combination of
>thresholds. The correct syntax for this is shown below:
>
> raw_thresh = >=75&&<100;
>
>That selects all values between 75 and 100.
>
>This is mentioned on page 44 of the MET users's guide:
>
>http://www.dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v5.1.pdf
>
>Lastly, I'm wondering why you're choosing a raw threshold for MODE at
all.
>Typically users do not need to specify a raw_thresh value. All it
does is
>set any values which do not meet that threshold criteria to a value
of 0.
>
>For example, we ran MODE to compare a forecast of reflectivity to
observed
>reflectivity. However, the forecasts were only for reflectivity
values of
>35 dBZ and higher while the observed reflectivity contained a full
range of
>values. Using MODE to compare forecast areas of reflectivity 35 dBZ
and
>higher to a full range of reflectivity values isn't really "fair".
The
>convolution, or smoothing, step will lead to smaller forecast objects
since
>the 35+ values are being smoothed with surrounding values of 0.
Therefore,
>we set the observed raw_thresh to >=35. That made it so that both
the
>forecast and observation fields had values of 35 dBz surrounded by
values
>of 0.
>
>Unless you're in a similar situation, where the range of values in
the
>forecast and observation field differ, I doubt the raw_thresh is the
>correct thing to be using.
>
>Hope that helps clarify.
>
>Thanks,
>John Halley Gotway
>
>On Tue, Apr 12, 2016 at 6:20 AM, Zhuang from Nuist via RT
<met_help at ucar.edu
>> wrote:
>
>>
>> Tue Apr 12 06:20:55 2016: Request 75882 was acted upon.
>> Transaction: Ticket created by zxrxz3212009 at 163.com
>> Queue: met_help
>> Subject: hello,sir. qustion about threshold in modeconfig file
>> Owner: Nobody
>> Requestors: zxrxz3212009 at 163.com
>> Status: new
>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
>>
>>
>> Hello,sir:
>> In MODEConfig file I want set raw threshld as "75<=thre
<100",
>> how would I set in MODEConfig file. I've tried .And there came out
a
>> syntax error.
>> Thank u very much and i'm looking forward to your early
reply.
>>
>
------------------------------------------------
Subject: hello,sir. qustion about threshold in modeconfig file
From: John Halley Gotway
Time: Tue Apr 12 21:09:24 2016
Unfortunately no, MET does not yet support percentile thresholds.
However,
we do plan to add support for it in the next major release of the
software.
Please note that in MODE there are two thresholds... the raw_thresh
and the
conv_thresh. The conv_thresh is what's used to actually define
objects in
MODE. As I mentioned, the majority of MODE users do not need to set
the
raw_thresh and instead use the conv_thresh to actually define objects.
Thanks,
John
On Tue, Apr 12, 2016 at 7:46 PM, Zhuang from Nuist via RT
<met_help at ucar.edu
> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
>
> Dir sir:
> Thanks very much for your kind consideration.
> As your said, compare objects with a fixed threshold is not
"fair",
> that's to say, if I want to compare the mesoscale system in a rain
case(for
> eg. >=50mm) is not resonable. But references like Ebert(2008);
> William(2009) all chosed a fixed threshold in order to compare the
object,
> like CRA method.
> Also, can mode use a percentile threshold instead of a fixed
one?
> Thanks very much!
>
>
>
>
>
>
>
>
> At 2016-04-13 00:00:19, "John Halley Gotway via RT"
<met_help at ucar.edu>
> wrote:
> >Hello, I see you have question about specifying a combination of
> >thresholds. The correct syntax for this is shown below:
> >
> > raw_thresh = >=75&&<100;
> >
> >That selects all values between 75 and 100.
> >
> >This is mentioned on page 44 of the MET users's guide:
> >
> >
>
http://www.dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v5.1.pdf
> >
> >Lastly, I'm wondering why you're choosing a raw threshold for MODE
at all.
> >Typically users do not need to specify a raw_thresh value. All it
does is
> >set any values which do not meet that threshold criteria to a value
of 0.
> >
> >For example, we ran MODE to compare a forecast of reflectivity to
observed
> >reflectivity. However, the forecasts were only for reflectivity
values of
> >35 dBZ and higher while the observed reflectivity contained a full
range
> of
> >values. Using MODE to compare forecast areas of reflectivity 35
dBZ and
> >higher to a full range of reflectivity values isn't really "fair".
The
> >convolution, or smoothing, step will lead to smaller forecast
objects
> since
> >the 35+ values are being smoothed with surrounding values of 0.
> Therefore,
> >we set the observed raw_thresh to >=35. That made it so that both
the
> >forecast and observation fields had values of 35 dBz surrounded by
values
> >of 0.
> >
> >Unless you're in a similar situation, where the range of values in
the
> >forecast and observation field differ, I doubt the raw_thresh is
the
> >correct thing to be using.
> >
> >Hope that helps clarify.
> >
> >Thanks,
> >John Halley Gotway
> >
> >On Tue, Apr 12, 2016 at 6:20 AM, Zhuang from Nuist via RT <
> met_help at ucar.edu
> >> wrote:
> >
> >>
> >> Tue Apr 12 06:20:55 2016: Request 75882 was acted upon.
> >> Transaction: Ticket created by zxrxz3212009 at 163.com
> >> Queue: met_help
> >> Subject: hello,sir. qustion about threshold in modeconfig
file
> >> Owner: Nobody
> >> Requestors: zxrxz3212009 at 163.com
> >> Status: new
> >> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
> >>
> >>
> >> Hello,sir:
> >> In MODEConfig file I want set raw threshld as "75<=thre
<100",
> >> how would I set in MODEConfig file. I've tried .And there came
out a
> >> syntax error.
> >> Thank u very much and i'm looking forward to your early
reply.
> >>
> >
>
>
------------------------------------------------
Subject: hello,sir. qustion about threshold in modeconfig file
From: Zhuang from Nuist
Time: Wed Apr 13 20:10:37 2016
Thank u very much, sir.
Another question, while using mtd(Mode Time Domain), fcst files
are ctl01.grb,...,ctl08.grb, out put from UPPV2.2.
My original wrfout file is wrfout_d02_2014-05-22_12:00:00 which
contain 25 times from 2014-05-22_12:00:00 to 2014-05-23_12:00:00. As I
want to use mtd to analyse the 3-hr accumulated precipitation, (1)
first I use NCO to sepetate original wrfout_d02_2014-05-22_12:00:00 to
8 files: wrfout_d02_2014-05-22_12:00:00, wrfout_d02_2014-05-
22_15:00:00,...,wrfout_d02_2014-05-23_09:00:00, each file contain only
one time. (2) then I ran a NCL file to caculate and write each 3-hr
accumulated precipitation(eg. wrfout_d02_2014-05-22_12:00:00 to
wrfout_d02_2014-05-22_15:00:00.....) to RAINC in each of the 8 files.
(3) I use UPPV2.2 to translate the 8 files above to .grb so I can use
mtd in METv5.1 to handle them, but as I see, the ACPCP are not the
same in 8 files, in ctl01.grb the var is named ACPCP and A00, but in
ctl01.grb it named ACPCP and A03, so I have no idea how to change the
MTDConfig file.
grid_res = 3;
////////////////////////////////////////////////////////////////////////////////
//
// Forecast and observation fields to be verified
//
fcst = {
field = {
name = "ACPCP";
level = "A00" ;
}
conv_radius = 36.0/grid_res; // in grid squares
conv_thresh = >=10.0;
}
obs = fcst;
I'm looking forward to your early reply!
At 2016-04-13 11:09:24, "John Halley Gotway via RT"
<met_help at ucar.edu> wrote:
>Unfortunately no, MET does not yet support percentile thresholds.
However,
>we do plan to add support for it in the next major release of the
software.
>
>Please note that in MODE there are two thresholds... the raw_thresh
and the
>conv_thresh. The conv_thresh is what's used to actually define
objects in
>MODE. As I mentioned, the majority of MODE users do not need to set
the
>raw_thresh and instead use the conv_thresh to actually define
objects.
>
>Thanks,
>John
>
>On Tue, Apr 12, 2016 at 7:46 PM, Zhuang from Nuist via RT
<met_help at ucar.edu
>> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
>>
>> Dir sir:
>> Thanks very much for your kind consideration.
>> As your said, compare objects with a fixed threshold is not
"fair",
>> that's to say, if I want to compare the mesoscale system in a rain
case(for
>> eg. >=50mm) is not resonable. But references like Ebert(2008);
>> William(2009) all chosed a fixed threshold in order to compare the
object,
>> like CRA method.
>> Also, can mode use a percentile threshold instead of a fixed
one?
>> Thanks very much!
>>
>>
>>
>>
>>
>>
>>
>>
>> At 2016-04-13 00:00:19, "John Halley Gotway via RT"
<met_help at ucar.edu>
>> wrote:
>> >Hello, I see you have question about specifying a combination of
>> >thresholds. The correct syntax for this is shown below:
>> >
>> > raw_thresh = >=75&&<100;
>> >
>> >That selects all values between 75 and 100.
>> >
>> >This is mentioned on page 44 of the MET users's guide:
>> >
>> >
>>
http://www.dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v5.1.pdf
>> >
>> >Lastly, I'm wondering why you're choosing a raw threshold for MODE
at all.
>> >Typically users do not need to specify a raw_thresh value. All it
does is
>> >set any values which do not meet that threshold criteria to a
value of 0.
>> >
>> >For example, we ran MODE to compare a forecast of reflectivity to
observed
>> >reflectivity. However, the forecasts were only for reflectivity
values of
>> >35 dBZ and higher while the observed reflectivity contained a full
range
>> of
>> >values. Using MODE to compare forecast areas of reflectivity 35
dBZ and
>> >higher to a full range of reflectivity values isn't really "fair".
The
>> >convolution, or smoothing, step will lead to smaller forecast
objects
>> since
>> >the 35+ values are being smoothed with surrounding values of 0.
>> Therefore,
>> >we set the observed raw_thresh to >=35. That made it so that both
the
>> >forecast and observation fields had values of 35 dBz surrounded by
values
>> >of 0.
>> >
>> >Unless you're in a similar situation, where the range of values in
the
>> >forecast and observation field differ, I doubt the raw_thresh is
the
>> >correct thing to be using.
>> >
>> >Hope that helps clarify.
>> >
>> >Thanks,
>> >John Halley Gotway
>> >
>> >On Tue, Apr 12, 2016 at 6:20 AM, Zhuang from Nuist via RT <
>> met_help at ucar.edu
>> >> wrote:
>> >
>> >>
>> >> Tue Apr 12 06:20:55 2016: Request 75882 was acted upon.
>> >> Transaction: Ticket created by zxrxz3212009 at 163.com
>> >> Queue: met_help
>> >> Subject: hello,sir. qustion about threshold in modeconfig
file
>> >> Owner: Nobody
>> >> Requestors: zxrxz3212009 at 163.com
>> >> Status: new
>> >> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
>> >>
>> >>
>> >> Hello,sir:
>> >> In MODEConfig file I want set raw threshld as
"75<=thre <100",
>> >> how would I set in MODEConfig file. I've tried .And there came
out a
>> >> syntax error.
>> >> Thank u very much and i'm looking forward to your early
reply.
>> >>
>> >
>>
>>
>
------------------------------------------------
Subject: hello,sir. qustion about threshold in modeconfig file
From: John Halley Gotway
Time: Mon Apr 18 10:11:39 2016
Hello, I see you have a question about running the MODE time domain
tool on
3-hourly accumulated precipitation. I read the description of your
processing and understand that you end up with 3-hourly precip for
most
output times but 0-hour accumulation for one of them.
I suspect that you have a 0-hour accumulation for the first output
time
(i.e. forecast hour 0) and 3-hourly accumulations for all other lead
times.
That makes complete sense. At forecast hour 0, you should of course
have
no accumulated precipitation. This is very simple to fix. When you
run
MODE Time Domain, you use the "-fcst" and "-obs" command line
arguments to
pass it the list of files to process. All you need to do is remove
the
0-hour files from that list.
Make sense?
Thanks,
John Halley Gotway
met_help at ucar.edu
On Wed, Apr 13, 2016 at 8:10 PM, Zhuang from Nuist via RT
<met_help at ucar.edu
> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
>
> Thank u very much, sir.
> Another question, while using mtd(Mode Time Domain), fcst files
are
> ctl01.grb,...,ctl08.grb, out put from UPPV2.2.
> My original wrfout file is wrfout_d02_2014-05-22_12:00:00 which
> contain 25 times from 2014-05-22_12:00:00 to 2014-05-23_12:00:00. As
I want
> to use mtd to analyse the 3-hr accumulated precipitation, (1) first
I use
> NCO to sepetate original wrfout_d02_2014-05-22_12:00:00 to 8 files:
> wrfout_d02_2014-05-22_12:00:00,
> wrfout_d02_2014-05-22_15:00:00,...,wrfout_d02_2014-05-23_09:00:00,
each
> file contain only one time. (2) then I ran a NCL file to caculate
and write
> each 3-hr accumulated precipitation(eg. wrfout_d02_2014-05-
22_12:00:00 to
> wrfout_d02_2014-05-22_15:00:00.....) to RAINC in each of the 8
files. (3)
> I use UPPV2.2 to translate the 8 files above to .grb so I can use
mtd in
> METv5.1 to handle them, but as I see, the ACPCP are not the same in
8
> files, in ctl01.grb the var is named ACPCP and A00, but in ctl01.grb
it
> named ACPCP and A03, so I have no idea how to change the MTDConfig
file.
>
>
> grid_res = 3;
>
>
>
>
////////////////////////////////////////////////////////////////////////////////
>
>
> //
> // Forecast and observation fields to be verified
> //
>
>
> fcst = {
> field = {
> name = "ACPCP";
> level = "A00" ;
> }
>
>
> conv_radius = 36.0/grid_res; // in grid squares
> conv_thresh = >=10.0;
>
>
> }
>
>
> obs = fcst;
>
>
> I'm looking forward to your early reply!
>
>
>
>
>
>
>
>
> At 2016-04-13 11:09:24, "John Halley Gotway via RT"
<met_help at ucar.edu>
> wrote:
> >Unfortunately no, MET does not yet support percentile thresholds.
> However,
> >we do plan to add support for it in the next major release of the
> software.
> >
> >Please note that in MODE there are two thresholds... the raw_thresh
and
> the
> >conv_thresh. The conv_thresh is what's used to actually define
objects in
> >MODE. As I mentioned, the majority of MODE users do not need to
set the
> >raw_thresh and instead use the conv_thresh to actually define
objects.
> >
> >Thanks,
> >John
> >
> >On Tue, Apr 12, 2016 at 7:46 PM, Zhuang from Nuist via RT <
> met_help at ucar.edu
> >> wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
> >>
> >> Dir sir:
> >> Thanks very much for your kind consideration.
> >> As your said, compare objects with a fixed threshold is not
> "fair",
> >> that's to say, if I want to compare the mesoscale system in a
rain
> case(for
> >> eg. >=50mm) is not resonable. But references like Ebert(2008);
> >> William(2009) all chosed a fixed threshold in order to compare
the
> object,
> >> like CRA method.
> >> Also, can mode use a percentile threshold instead of a
fixed one?
> >> Thanks very much!
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >> At 2016-04-13 00:00:19, "John Halley Gotway via RT"
<met_help at ucar.edu>
> >> wrote:
> >> >Hello, I see you have question about specifying a combination of
> >> >thresholds. The correct syntax for this is shown below:
> >> >
> >> > raw_thresh = >=75&&<100;
> >> >
> >> >That selects all values between 75 and 100.
> >> >
> >> >This is mentioned on page 44 of the MET users's guide:
> >> >
> >> >
> >>
>
http://www.dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v5.1.pdf
> >> >
> >> >Lastly, I'm wondering why you're choosing a raw threshold for
MODE at
> all.
> >> >Typically users do not need to specify a raw_thresh value. All
it
> does is
> >> >set any values which do not meet that threshold criteria to a
value of
> 0.
> >> >
> >> >For example, we ran MODE to compare a forecast of reflectivity
to
> observed
> >> >reflectivity. However, the forecasts were only for reflectivity
> values of
> >> >35 dBZ and higher while the observed reflectivity contained a
full
> range
> >> of
> >> >values. Using MODE to compare forecast areas of reflectivity 35
dBZ
> and
> >> >higher to a full range of reflectivity values isn't really
"fair". The
> >> >convolution, or smoothing, step will lead to smaller forecast
objects
> >> since
> >> >the 35+ values are being smoothed with surrounding values of 0.
> >> Therefore,
> >> >we set the observed raw_thresh to >=35. That made it so that
both the
> >> >forecast and observation fields had values of 35 dBz surrounded
by
> values
> >> >of 0.
> >> >
> >> >Unless you're in a similar situation, where the range of values
in the
> >> >forecast and observation field differ, I doubt the raw_thresh is
the
> >> >correct thing to be using.
> >> >
> >> >Hope that helps clarify.
> >> >
> >> >Thanks,
> >> >John Halley Gotway
> >> >
> >> >On Tue, Apr 12, 2016 at 6:20 AM, Zhuang from Nuist via RT <
> >> met_help at ucar.edu
> >> >> wrote:
> >> >
> >> >>
> >> >> Tue Apr 12 06:20:55 2016: Request 75882 was acted upon.
> >> >> Transaction: Ticket created by zxrxz3212009 at 163.com
> >> >> Queue: met_help
> >> >> Subject: hello,sir. qustion about threshold in modeconfig
file
> >> >> Owner: Nobody
> >> >> Requestors: zxrxz3212009 at 163.com
> >> >> Status: new
> >> >> Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=75882 >
> >> >>
> >> >>
> >> >> Hello,sir:
> >> >> In MODEConfig file I want set raw threshld as
"75<=thre
> <100",
> >> >> how would I set in MODEConfig file. I've tried .And there
came out a
> >> >> syntax error.
> >> >> Thank u very much and i'm looking forward to your early
reply.
> >> >>
> >> >
> >>
> >>
> >
>
>
------------------------------------------------
More information about the Met_help
mailing list