[Met_help] [rt.rap.ucar.edu #96074] History for wind_thresh

Thu Oct 1 14:32:17 MDT 2020

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

Hello,

I have a question about setting up wind_thresh in config file: because the
wind can be negative and positive, so how to set them correctly?
e.g, which is correct:
wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10 ||>=-10&&<=-5,
>=10.0||<=-10 ];
or
|wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];

Thank you.
Binyu

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

Subject: wind_thresh
From: John Halley Gotway
Time: Fri Jul 31 14:07:41 2020

Hi Binyu,

The wind_thresh is a filtering threshold which is applied to the wind
SPEED. While the individual U and V components can be positive or
negative,
the wind speed is always >= 0.

The wind_thresh applies to the processing of the VL1L2, VAL1L2, and
VCNT
line types which process U/V pairs. Let's say you've set:
wind_thresh = [ NA, ge1, ge5, ge10 ];

Instead of seeing only 1 VL1L2 output line, you'll get 4... one for
each
wind_thresh chosen.
The first one (wind_thresh = NA) will include all U/V pairs in the
masking
area.
The second one (wind_thresh ge1) will only include those U/V pairs
whose
corresponding wind speed is at least 1 m/s.
The third and fourth ones (wind_thresh ge5 and ge10) will only include
those U/V pairs whose corresponding wind speed is at least 5 m/s or
10m/s,
respectively.

The idea here is checking to see how the model performs light and
variable
winds (i.e. low wind speed) versus well-defined winds (i.e. high wind
speed).

Thanks,
John

On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> Transaction: Ticket created by binyu.wang at noaa.gov
>        Queue: met_help
>      Subject: wind_thresh
>        Owner: Nobody
>   Requestors: binyu.wang at noaa.gov
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
>
> Hello,
>
> I have a question about setting up wind_thresh in config file:
because the
> wind can be negative and positive, so how to set them correctly?
> e.g, which is correct:
> wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10 ||>=-10&&<=-5,
> >=10.0||<=-10 ];
> or
> |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
>
> Thank you.
> Binyu
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Fri Jul 31 15:33:58 2020

John,

That makes sense for wind SPEED. Please look at the point_stat config:

/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
PointStatConfig

field = [

     {

        name       = "UGRD";

        level      = [ "P250" ];

        cat_thresh = [ >=1.0, >=5.0 ];

      },

the cat_thresh ONLY has positive too, does that mean both directions
will
be considered? (>= refers either ge 5 or le -5)?

One more question here: following your suggestion, I did create VL1L2
and
VCNT,but I got strange error (
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
when I
tried to load to METviewer ( I mentioned that in one of my previous
email,
but didn't get reply). Is there any plot in METviwer that can compare
the
wind direction b/t obs. and forecast?

Thank you.

Binyu

On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Hi Binyu,
>
> The wind_thresh is a filtering threshold which is applied to the
wind
> SPEED. While the individual U and V components can be positive or
negative,
> the wind speed is always >= 0.
>
> The wind_thresh applies to the processing of the VL1L2, VAL1L2, and
VCNT
> line types which process U/V pairs. Let's say you've set:
> wind_thresh = [ NA, ge1, ge5, ge10 ];
>
> Instead of seeing only 1 VL1L2 output line, you'll get 4... one for
each
> wind_thresh chosen.
> The first one (wind_thresh = NA) will include all U/V pairs in the
masking
> area.
> The second one (wind_thresh ge1) will only include those U/V pairs
whose
> corresponding wind speed is at least 1 m/s.
> The third and fourth ones (wind_thresh ge5 and ge10) will only
include
> those U/V pairs whose corresponding wind speed is at least 5 m/s or
10m/s,
> respectively.
>
> The idea here is checking to see how the model performs light and
variable
> winds (i.e. low wind speed) versus well-defined winds (i.e. high
wind
> speed).
>
> Thanks,
> John
>
> On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
> >
> > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> > Transaction: Ticket created by binyu.wang at noaa.gov
> >        Queue: met_help
> >      Subject: wind_thresh
> >        Owner: Nobody
> >   Requestors: binyu.wang at noaa.gov
> >       Status: new
> >  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> >
> > Hello,
> >
> > I have a question about setting up wind_thresh in config file:
because
> the
> > wind can be negative and positive, so how to set them correctly?
> > e.g, which is correct:
> > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10 ||>=-10&&<=-
5,
> > >=10.0||<=-10 ];
> > or
> > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> >
> > Thank you.
> > Binyu
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Fri Jul 31 16:10:38 2020

Binyu,

You're right, the cat_thresh for UGRD of >=1.0 would only threshold
the
positive values. If you want to apply both positive and negative you
could
use:

cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];

Table 7.19 in the MET User's Guide describes the contents of the VCNT
line
type:
[image: Screen Shot 2020-07-31 at 4.06.47 PM.png]

Looking at the description of those statistics, it looks like you may
be
interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and DIR_ABSERR. And
METviewer should be able to plot those statistics from the VCNT line
type.
That being said, I have never actually done this myself for a model
evaluation.

If METviewer isn't loading VCNT lines or plotting the statistics from
them,
we'll need to let Tatiana Burek know so that she can fix it.

Thanks,
John

On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> John,
>
> That makes sense for wind SPEED. Please look at the point_stat
config:
>
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> PointStatConfig
>
> field = [
>
>      {
>
>         name       = "UGRD";
>
>         level      = [ "P250" ];
>
>         cat_thresh = [ >=1.0, >=5.0 ];
>
>       },
>
>
> the cat_thresh ONLY has positive too, does that mean both directions
will
> be considered? (>= refers either ge 5 or le -5)?
>
>
> One more question here: following your suggestion, I did create
VL1L2 and
> VCNT,but I got strange error (
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
when I
> tried to load to METviewer ( I mentioned that in one of my previous
email,
> but didn't get reply). Is there any plot in METviwer that can
compare the
> wind direction b/t obs. and forecast?
>
>
> Thank you.
>
> Binyu
>
>
>
>
> On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Hi Binyu,
> >
> > The wind_thresh is a filtering threshold which is applied to the
wind
> > SPEED. While the individual U and V components can be positive or
> negative,
> > the wind speed is always >= 0.
> >
> > The wind_thresh applies to the processing of the VL1L2, VAL1L2,
and VCNT
> > line types which process U/V pairs. Let's say you've set:
> > wind_thresh = [ NA, ge1, ge5, ge10 ];
> >
> > Instead of seeing only 1 VL1L2 output line, you'll get 4... one
for each
> > wind_thresh chosen.
> > The first one (wind_thresh = NA) will include all U/V pairs in the
> masking
> > area.
> > The second one (wind_thresh ge1) will only include those U/V pairs
whose
> > corresponding wind speed is at least 1 m/s.
> > The third and fourth ones (wind_thresh ge5 and ge10) will only
include
> > those U/V pairs whose corresponding wind speed is at least 5 m/s
or
> 10m/s,
> > respectively.
> >
> > The idea here is checking to see how the model performs light and
> variable
> > winds (i.e. low wind speed) versus well-defined winds (i.e. high
wind
> > speed).
> >
> > Thanks,
> > John
> >
> > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> > > Transaction: Ticket created by binyu.wang at noaa.gov
> > >        Queue: met_help
> > >      Subject: wind_thresh
> > >        Owner: Nobody
> > >   Requestors: binyu.wang at noaa.gov
> > >       Status: new
> > >  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >
> > >
> > >
> > > Hello,
> > >
> > > I have a question about setting up wind_thresh in config file:
because
> > the
> > > wind can be negative and positive, so how to set them correctly?
> > > e.g, which is correct:
> > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10 ||>=-
10&&<=-5,
> > > >=10.0||<=-10 ];
> > > or
> > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > >
> > > Thank you.
> > > Binyu
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Fri Jul 31 16:44:54 2020

John,

Should I send another ticket to "Tatiana" about my log error to upload
point_stat*stat files to METViwer?

I just tried the new config file (cat_thresh = [ <=-1.0||>=1.0,
<=-5.0||>=5.0 ];) and the output are the SAME as using the old config
file (
 cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few matching
pairs
(only one or two pairs)?

Binyu

On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Binyu,
>
> You're right, the cat_thresh for UGRD of >=1.0 would only threshold
the
> positive values. If you want to apply both positive and negative you
could
> use:
>
> cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
>
> Table 7.19 in the MET User's Guide describes the contents of the
VCNT line
> type:
> [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
>
> Looking at the description of those statistics, it looks like you
may be
> interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and DIR_ABSERR. And
> METviewer should be able to plot those statistics from the VCNT line
type.
> That being said, I have never actually done this myself for a model
> evaluation.
>
> If METviewer isn't loading VCNT lines or plotting the statistics
from them,
> we'll need to let Tatiana Burek know so that she can fix it.
>
> Thanks,
> John
>
>
>
> On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> > John,
> >
> > That makes sense for wind SPEED. Please look at the point_stat
config:
> >
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > PointStatConfig
> >
> > field = [
> >
> >      {
> >
> >         name       = "UGRD";
> >
> >         level      = [ "P250" ];
> >
> >         cat_thresh = [ >=1.0, >=5.0 ];
> >
> >       },
> >
> >
> > the cat_thresh ONLY has positive too, does that mean both
directions will
> > be considered? (>= refers either ge 5 or le -5)?
> >
> >
> > One more question here: following your suggestion, I did create
VL1L2 and
> > VCNT,but I got strange error (
> > /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
when I
> > tried to load to METviewer ( I mentioned that in one of my
previous
> email,
> > but didn't get reply). Is there any plot in METviwer that can
compare the
> > wind direction b/t obs. and forecast?
> >
> >
> > Thank you.
> >
> > Binyu
> >
> >
> >
> >
> > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Hi Binyu,
> > >
> > > The wind_thresh is a filtering threshold which is applied to the
wind
> > > SPEED. While the individual U and V components can be positive
or
> > negative,
> > > the wind speed is always >= 0.
> > >
> > > The wind_thresh applies to the processing of the VL1L2, VAL1L2,
and
> VCNT
> > > line types which process U/V pairs. Let's say you've set:
> > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > >
> > > Instead of seeing only 1 VL1L2 output line, you'll get 4... one
for
> each
> > > wind_thresh chosen.
> > > The first one (wind_thresh = NA) will include all U/V pairs in
the
> > masking
> > > area.
> > > The second one (wind_thresh ge1) will only include those U/V
pairs
> whose
> > > corresponding wind speed is at least 1 m/s.
> > > The third and fourth ones (wind_thresh ge5 and ge10) will only
include
> > > those U/V pairs whose corresponding wind speed is at least 5 m/s
or
> > 10m/s,
> > > respectively.
> > >
> > > The idea here is checking to see how the model performs light
and
> > variable
> > > winds (i.e. low wind speed) versus well-defined winds (i.e. high
wind
> > > speed).
> > >
> > > Thanks,
> > > John
> > >
> > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > > >        Queue: met_help
> > > >      Subject: wind_thresh
> > > >        Owner: Nobody
> > > >   Requestors: binyu.wang at noaa.gov
> > > >       Status: new
> > > >  Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > >
> > > >
> > > >
> > > > Hello,
> > > >
> > > > I have a question about setting up wind_thresh in config file:
> because
> > > the
> > > > wind can be negative and positive, so how to set them
correctly?
> > > > e.g, which is correct:
> > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10 ||>=-
10&&<=-5,
> > > > >=10.0||<=-10 ];
> > > > or
> > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > > >
> > > > Thank you.
> > > > Binyu
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Fri Jul 31 16:52:28 2020

Binyu,

I'm guessing that the output is not actually the same. In the VL1L2
and
VCNT lines, the FCST_THRESH and OBS_THRESH columns should now indicate
that
you're using different thresholds: '>=1.0' should now be changed to
'<=-1.0||>=1.0'

You should confirm that those have changed to be confident that you
actually ran the modified configuration. Once you've confirmed that,
if the
counts and stats are still the same, that's really just a function of
your
data... about which I have no idea.

John

On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> John,
>
> Should I send another ticket to "Tatiana" about my log error to
upload
> point_stat*stat files to METViwer?
>
> I just tried the new config file (cat_thresh = [ <=-1.0||>=1.0,
> <=-5.0||>=5.0 ];) and the output are the SAME as using the old
config file
> (
>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few
matching pairs
> (only one or two pairs)?
>
> Binyu
>
> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Binyu,
> >
> > You're right, the cat_thresh for UGRD of >=1.0 would only
threshold the
> > positive values. If you want to apply both positive and negative
you
> could
> > use:
> >
> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> >
> > Table 7.19 in the MET User's Guide describes the contents of the
VCNT
> line
> > type:
> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> >
> > Looking at the description of those statistics, it looks like you
may be
> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and DIR_ABSERR. And
> > METviewer should be able to plot those statistics from the VCNT
line
> type.
> > That being said, I have never actually done this myself for a
model
> > evaluation.
> >
> > If METviewer isn't loading VCNT lines or plotting the statistics
from
> them,
> > we'll need to let Tatiana Burek know so that she can fix it.
> >
> > Thanks,
> > John
> >
> >
> >
> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > >
> > > John,
> > >
> > > That makes sense for wind SPEED. Please look at the point_stat
config:
> > >
> > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > PointStatConfig
> > >
> > > field = [
> > >
> > >      {
> > >
> > >         name       = "UGRD";
> > >
> > >         level      = [ "P250" ];
> > >
> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > >
> > >       },
> > >
> > >
> > > the cat_thresh ONLY has positive too, does that mean both
directions
> will
> > > be considered? (>= refers either ge 5 or le -5)?
> > >
> > >
> > > One more question here: following your suggestion, I did create
VL1L2
> and
> > > VCNT,but I got strange error (
> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> when I
> > > tried to load to METviewer ( I mentioned that in one of my
previous
> > email,
> > > but didn't get reply). Is there any plot in METviwer that can
compare
> the
> > > wind direction b/t obs. and forecast?
> > >
> > >
> > > Thank you.
> > >
> > > Binyu
> > >
> > >
> > >
> > >
> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > > Hi Binyu,
> > > >
> > > > The wind_thresh is a filtering threshold which is applied to
the wind
> > > > SPEED. While the individual U and V components can be positive
or
> > > negative,
> > > > the wind speed is always >= 0.
> > > >
> > > > The wind_thresh applies to the processing of the VL1L2,
VAL1L2, and
> > VCNT
> > > > line types which process U/V pairs. Let's say you've set:
> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > >
> > > > Instead of seeing only 1 VL1L2 output line, you'll get 4...
one for
> > each
> > > > wind_thresh chosen.
> > > > The first one (wind_thresh = NA) will include all U/V pairs in
the
> > > masking
> > > > area.
> > > > The second one (wind_thresh ge1) will only include those U/V
pairs
> > whose
> > > > corresponding wind speed is at least 1 m/s.
> > > > The third and fourth ones (wind_thresh ge5 and ge10) will only
> include
> > > > those U/V pairs whose corresponding wind speed is at least 5
m/s or
> > > 10m/s,
> > > > respectively.
> > > >
> > > > The idea here is checking to see how the model performs light
and
> > > variable
> > > > winds (i.e. low wind speed) versus well-defined winds (i.e.
high wind
> > > > speed).
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > > > >        Queue: met_help
> > > > >      Subject: wind_thresh
> > > > >        Owner: Nobody
> > > > >   Requestors: binyu.wang at noaa.gov
> > > > >       Status: new
> > > > >  Ticket <URL:
> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > >
> > > > >
> > > > >
> > > > > Hello,
> > > > >
> > > > > I have a question about setting up wind_thresh in config
file:
> > because
> > > > the
> > > > > wind can be negative and positive, so how to set them
correctly?
> > > > > e.g, which is correct:
> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10
> ||>=-10&&<=-5,
> > > > > >=10.0||<=-10 ];
> > > > > or
> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > > > >
> > > > > Thank you.
> > > > > Binyu
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Fri Jul 31 16:58:31 2020

To keep these issues straight, I just created a separate RT ticket for
Tatiana to look at. Hopefully you received a copy of it.

Thanks,
John

On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <johnhg at ucar.edu>
wrote:

> Binyu,
>
> I'm guessing that the output is not actually the same. In the VL1L2
and
> VCNT lines, the FCST_THRESH and OBS_THRESH columns should now
indicate that
> you're using different thresholds: '>=1.0' should now be changed to
> '<=-1.0||>=1.0'
>
> You should confirm that those have changed to be confident that you
> actually ran the modified configuration. Once you've confirmed that,
if the
> counts and stats are still the same, that's really just a function
of your
> data... about which I have no idea.
>
> John
>
> On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>>
>> John,
>>
>> Should I send another ticket to "Tatiana" about my log error to
upload
>> point_stat*stat files to METViwer?
>>
>> I just tried the new config file (cat_thresh = [ <=-1.0||>=1.0,
>> <=-5.0||>=5.0 ];) and the output are the SAME as using the old
config
>> file (
>>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few
matching
>> pairs
>> (only one or two pairs)?
>>
>> Binyu
>>
>> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
>> met_help at ucar.edu>
>> wrote:
>>
>> > Binyu,
>> >
>> > You're right, the cat_thresh for UGRD of >=1.0 would only
threshold the
>> > positive values. If you want to apply both positive and negative
you
>> could
>> > use:
>> >
>> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
>> >
>> > Table 7.19 in the MET User's Guide describes the contents of the
VCNT
>> line
>> > type:
>> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
>> >
>> > Looking at the description of those statistics, it looks like you
may be
>> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and DIR_ABSERR.
And
>> > METviewer should be able to plot those statistics from the VCNT
line
>> type.
>> > That being said, I have never actually done this myself for a
model
>> > evaluation.
>> >
>> > If METviewer isn't loading VCNT lines or plotting the statistics
from
>> them,
>> > we'll need to let Tatiana Burek know so that she can fix it.
>> >
>> > Thanks,
>> > John
>> >
>> >
>> >
>> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT <
>> > met_help at ucar.edu> wrote:
>> >
>> > >
>> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>> > >
>> > > John,
>> > >
>> > > That makes sense for wind SPEED. Please look at the point_stat
config:
>> > >
>> > >
>> > >
>> >
>>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
>> > > PointStatConfig
>> > >
>> > > field = [
>> > >
>> > >      {
>> > >
>> > >         name       = "UGRD";
>> > >
>> > >         level      = [ "P250" ];
>> > >
>> > >         cat_thresh = [ >=1.0, >=5.0 ];
>> > >
>> > >       },
>> > >
>> > >
>> > > the cat_thresh ONLY has positive too, does that mean both
directions
>> will
>> > > be considered? (>= refers either ge 5 or le -5)?
>> > >
>> > >
>> > > One more question here: following your suggestion, I did create
VL1L2
>> and
>> > > VCNT,but I got strange error (
>> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
>> when I
>> > > tried to load to METviewer ( I mentioned that in one of my
previous
>> > email,
>> > > but didn't get reply). Is there any plot in METviwer that can
compare
>> the
>> > > wind direction b/t obs. and forecast?
>> > >
>> > >
>> > > Thank you.
>> > >
>> > > Binyu
>> > >
>> > >
>> > >
>> > >
>> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT <
>> > > met_help at ucar.edu>
>> > > wrote:
>> > >
>> > > > Hi Binyu,
>> > > >
>> > > > The wind_thresh is a filtering threshold which is applied to
the
>> wind
>> > > > SPEED. While the individual U and V components can be
positive or
>> > > negative,
>> > > > the wind speed is always >= 0.
>> > > >
>> > > > The wind_thresh applies to the processing of the VL1L2,
VAL1L2, and
>> > VCNT
>> > > > line types which process U/V pairs. Let's say you've set:
>> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
>> > > >
>> > > > Instead of seeing only 1 VL1L2 output line, you'll get 4...
one for
>> > each
>> > > > wind_thresh chosen.
>> > > > The first one (wind_thresh = NA) will include all U/V pairs
in the
>> > > masking
>> > > > area.
>> > > > The second one (wind_thresh ge1) will only include those U/V
pairs
>> > whose
>> > > > corresponding wind speed is at least 1 m/s.
>> > > > The third and fourth ones (wind_thresh ge5 and ge10) will
only
>> include
>> > > > those U/V pairs whose corresponding wind speed is at least 5
m/s or
>> > > 10m/s,
>> > > > respectively.
>> > > >
>> > > > The idea here is checking to see how the model performs light
and
>> > > variable
>> > > > winds (i.e. low wind speed) versus well-defined winds (i.e.
high
>> wind
>> > > > speed).
>> > > >
>> > > > Thanks,
>> > > > John
>> > > >
>> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT <
>> > > > met_help at ucar.edu> wrote:
>> > > >
>> > > > >
>> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
>> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
>> > > > >        Queue: met_help
>> > > > >      Subject: wind_thresh
>> > > > >        Owner: Nobody
>> > > > >   Requestors: binyu.wang at noaa.gov
>> > > > >       Status: new
>> > > > >  Ticket <URL:
>> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>> > > >
>> > > > >
>> > > > >
>> > > > > Hello,
>> > > > >
>> > > > > I have a question about setting up wind_thresh in config
file:
>> > because
>> > > > the
>> > > > > wind can be negative and positive, so how to set them
correctly?
>> > > > > e.g, which is correct:
>> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10
>> ||>=-10&&<=-5,
>> > > > > >=10.0||<=-10 ];
>> > > > > or
>> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
>> > > > >
>> > > > > Thank you.
>> > > > > Binyu
>> > > > >
>> > > > >
>> > > >
>> > > >
>> > >
>> > >
>> >
>> >
>>
>>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Fri Jul 31 17:45:27 2020

All the "FCST_THRESH and OBS_THRESH columns" are "NA". Does that
indicate
something wrong with my setting?

Here is my config file and one output:

/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
PointStatConfig

/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/ush/out/SA/point_stat/
point_stat_120000L_20200228_120000V_vl1l2.txt

Binyu

On Fri, Jul 31, 2020 at 6:52 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Binyu,
>
> I'm guessing that the output is not actually the same. In the VL1L2
and
> VCNT lines, the FCST_THRESH and OBS_THRESH columns should now
indicate that
> you're using different thresholds: '>=1.0' should now be changed to
> '<=-1.0||>=1.0'
>
> You should confirm that those have changed to be confident that you
> actually ran the modified configuration. Once you've confirmed that,
if the
> counts and stats are still the same, that's really just a function
of your
> data... about which I have no idea.
>
> John
>
> On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> > John,
> >
> > Should I send another ticket to "Tatiana" about my log error to
upload
> > point_stat*stat files to METViwer?
> >
> > I just tried the new config file (cat_thresh = [ <=-1.0||>=1.0,
> > <=-5.0||>=5.0 ];) and the output are the SAME as using the old
config
> file
> > (
> >  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few
matching
> pairs
> > (only one or two pairs)?
> >
> > Binyu
> >
> > On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Binyu,
> > >
> > > You're right, the cat_thresh for UGRD of >=1.0 would only
threshold the
> > > positive values. If you want to apply both positive and negative
you
> > could
> > > use:
> > >
> > > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > >
> > > Table 7.19 in the MET User's Guide describes the contents of the
VCNT
> > line
> > > type:
> > > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > >
> > > Looking at the description of those statistics, it looks like
you may
> be
> > > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and DIR_ABSERR.
And
> > > METviewer should be able to plot those statistics from the VCNT
line
> > type.
> > > That being said, I have never actually done this myself for a
model
> > > evaluation.
> > >
> > > If METviewer isn't loading VCNT lines or plotting the statistics
from
> > them,
> > > we'll need to let Tatiana Burek know so that she can fix it.
> > >
> > > Thanks,
> > > John
> > >
> > >
> > >
> > > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> > > >
> > > > John,
> > > >
> > > > That makes sense for wind SPEED. Please look at the point_stat
> config:
> > > >
> > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > PointStatConfig
> > > >
> > > > field = [
> > > >
> > > >      {
> > > >
> > > >         name       = "UGRD";
> > > >
> > > >         level      = [ "P250" ];
> > > >
> > > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > >
> > > >       },
> > > >
> > > >
> > > > the cat_thresh ONLY has positive too, does that mean both
directions
> > will
> > > > be considered? (>= refers either ge 5 or le -5)?
> > > >
> > > >
> > > > One more question here: following your suggestion, I did
create VL1L2
> > and
> > > > VCNT,but I got strange error (
> > > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > when I
> > > > tried to load to METviewer ( I mentioned that in one of my
previous
> > > email,
> > > > but didn't get reply). Is there any plot in METviwer that can
compare
> > the
> > > > wind direction b/t obs. and forecast?
> > > >
> > > >
> > > > Thank you.
> > > >
> > > > Binyu
> > > >
> > > >
> > > >
> > > >
> > > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > > Hi Binyu,
> > > > >
> > > > > The wind_thresh is a filtering threshold which is applied to
the
> wind
> > > > > SPEED. While the individual U and V components can be
positive or
> > > > negative,
> > > > > the wind speed is always >= 0.
> > > > >
> > > > > The wind_thresh applies to the processing of the VL1L2,
VAL1L2, and
> > > VCNT
> > > > > line types which process U/V pairs. Let's say you've set:
> > > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > >
> > > > > Instead of seeing only 1 VL1L2 output line, you'll get 4...
one for
> > > each
> > > > > wind_thresh chosen.
> > > > > The first one (wind_thresh = NA) will include all U/V pairs
in the
> > > > masking
> > > > > area.
> > > > > The second one (wind_thresh ge1) will only include those U/V
pairs
> > > whose
> > > > > corresponding wind speed is at least 1 m/s.
> > > > > The third and fourth ones (wind_thresh ge5 and ge10) will
only
> > include
> > > > > those U/V pairs whose corresponding wind speed is at least 5
m/s or
> > > > 10m/s,
> > > > > respectively.
> > > > >
> > > > > The idea here is checking to see how the model performs
light and
> > > > variable
> > > > > winds (i.e. low wind speed) versus well-defined winds (i.e.
high
> wind
> > > > > speed).
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > >
> > > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> > > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > > > > >        Queue: met_help
> > > > > >      Subject: wind_thresh
> > > > > >        Owner: Nobody
> > > > > >   Requestors: binyu.wang at noaa.gov
> > > > > >       Status: new
> > > > > >  Ticket <URL:
> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > >
> > > > > >
> > > > > >
> > > > > > Hello,
> > > > > >
> > > > > > I have a question about setting up wind_thresh in config
file:
> > > because
> > > > > the
> > > > > > wind can be negative and positive, so how to set them
correctly?
> > > > > > e.g, which is correct:
> > > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10
> > ||>=-10&&<=-5,
> > > > > > >=10.0||<=-10 ];
> > > > > > or
> > > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > > > > >
> > > > > > Thank you.
> > > > > > Binyu
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Thu Aug 06 13:30:27 2020

Hello again,

Following your suggestion, I tried to look for more sites for
verification
by using prepbufr data. Unfortunately I always got "0 matching pair"
when I
applied PB2NC and ensemble_stat. Here is the directory:

cd
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf

% ./test_bufr.ecf >& bufr.log3

 I tried to diagnose the problem by converting the prepbufr data to
grid
data using point_to_grid, the script is:
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh

One of the output is:

/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
regrid.prebufr.nc.txt

It seems the PRES field does not look correct to me: it is over
~20000mb,
which is impossible. I confirmed with prepbufr people, the unit inside
of
the prepbufr is mb, and the magnitude should be <1000mb. But I can not
find
where the error is. Any suggestions?

Thank you.
Binyu

On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> To keep these issues straight, I just created a separate RT ticket
for
> Tatiana to look at. Hopefully you received a copy of it.
>
> Thanks,
> John
>
> On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <johnhg at ucar.edu>
> wrote:
>
> > Binyu,
> >
> > I'm guessing that the output is not actually the same. In the
VL1L2 and
> > VCNT lines, the FCST_THRESH and OBS_THRESH columns should now
indicate
> that
> > you're using different thresholds: '>=1.0' should now be changed
to
> > '<=-1.0||>=1.0'
> >
> > You should confirm that those have changed to be confident that
you
> > actually ran the modified configuration. Once you've confirmed
that, if
> the
> > counts and stats are still the same, that's really just a function
of
> your
> > data... about which I have no idea.
> >
> > John
> >
> > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu> wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >>
> >> John,
> >>
> >> Should I send another ticket to "Tatiana" about my log error to
upload
> >> point_stat*stat files to METViwer?
> >>
> >> I just tried the new config file (cat_thresh = [ <=-1.0||>=1.0,
> >> <=-5.0||>=5.0 ];) and the output are the SAME as using the old
config
> >> file (
> >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few
matching
> >> pairs
> >> (only one or two pairs)?
> >>
> >> Binyu
> >>
> >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> >> met_help at ucar.edu>
> >> wrote:
> >>
> >> > Binyu,
> >> >
> >> > You're right, the cat_thresh for UGRD of >=1.0 would only
threshold
> the
> >> > positive values. If you want to apply both positive and
negative you
> >> could
> >> > use:
> >> >
> >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> >> >
> >> > Table 7.19 in the MET User's Guide describes the contents of
the VCNT
> >> line
> >> > type:
> >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> >> >
> >> > Looking at the description of those statistics, it looks like
you may
> be
> >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and DIR_ABSERR.
And
> >> > METviewer should be able to plot those statistics from the VCNT
line
> >> type.
> >> > That being said, I have never actually done this myself for a
model
> >> > evaluation.
> >> >
> >> > If METviewer isn't loading VCNT lines or plotting the
statistics from
> >> them,
> >> > we'll need to let Tatiana Burek know so that she can fix it.
> >> >
> >> > Thanks,
> >> > John
> >> >
> >> >
> >> >
> >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT <
> >> > met_help at ucar.edu> wrote:
> >> >
> >> > >
> >> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> >> > >
> >> > > John,
> >> > >
> >> > > That makes sense for wind SPEED. Please look at the
point_stat
> config:
> >> > >
> >> > >
> >> > >
> >> >
> >>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> >> > > PointStatConfig
> >> > >
> >> > > field = [
> >> > >
> >> > >      {
> >> > >
> >> > >         name       = "UGRD";
> >> > >
> >> > >         level      = [ "P250" ];
> >> > >
> >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> >> > >
> >> > >       },
> >> > >
> >> > >
> >> > > the cat_thresh ONLY has positive too, does that mean both
directions
> >> will
> >> > > be considered? (>= refers either ge 5 or le -5)?
> >> > >
> >> > >
> >> > > One more question here: following your suggestion, I did
create
> VL1L2
> >> and
> >> > > VCNT,but I got strange error (
> >> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> >> when I
> >> > > tried to load to METviewer ( I mentioned that in one of my
previous
> >> > email,
> >> > > but didn't get reply). Is there any plot in METviwer that can
> compare
> >> the
> >> > > wind direction b/t obs. and forecast?
> >> > >
> >> > >
> >> > > Thank you.
> >> > >
> >> > > Binyu
> >> > >
> >> > >
> >> > >
> >> > >
> >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT <
> >> > > met_help at ucar.edu>
> >> > > wrote:
> >> > >
> >> > > > Hi Binyu,
> >> > > >
> >> > > > The wind_thresh is a filtering threshold which is applied
to the
> >> wind
> >> > > > SPEED. While the individual U and V components can be
positive or
> >> > > negative,
> >> > > > the wind speed is always >= 0.
> >> > > >
> >> > > > The wind_thresh applies to the processing of the VL1L2,
VAL1L2,
> and
> >> > VCNT
> >> > > > line types which process U/V pairs. Let's say you've set:
> >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> >> > > >
> >> > > > Instead of seeing only 1 VL1L2 output line, you'll get 4...
one
> for
> >> > each
> >> > > > wind_thresh chosen.
> >> > > > The first one (wind_thresh = NA) will include all U/V pairs
in the
> >> > > masking
> >> > > > area.
> >> > > > The second one (wind_thresh ge1) will only include those
U/V pairs
> >> > whose
> >> > > > corresponding wind speed is at least 1 m/s.
> >> > > > The third and fourth ones (wind_thresh ge5 and ge10) will
only
> >> include
> >> > > > those U/V pairs whose corresponding wind speed is at least
5 m/s
> or
> >> > > 10m/s,
> >> > > > respectively.
> >> > > >
> >> > > > The idea here is checking to see how the model performs
light and
> >> > > variable
> >> > > > winds (i.e. low wind speed) versus well-defined winds (i.e.
high
> >> wind
> >> > > > speed).
> >> > > >
> >> > > > Thanks,
> >> > > > John
> >> > > >
> >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via RT
<
> >> > > > met_help at ucar.edu> wrote:
> >> > > >
> >> > > > >
> >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> >> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> >> > > > >        Queue: met_help
> >> > > > >      Subject: wind_thresh
> >> > > > >        Owner: Nobody
> >> > > > >   Requestors: binyu.wang at noaa.gov
> >> > > > >       Status: new
> >> > > > >  Ticket <URL:
> >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >> > > >
> >> > > > >
> >> > > > >
> >> > > > > Hello,
> >> > > > >
> >> > > > > I have a question about setting up wind_thresh in config
file:
> >> > because
> >> > > > the
> >> > > > > wind can be negative and positive, so how to set them
correctly?
> >> > > > > e.g, which is correct:
> >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10
> >> ||>=-10&&<=-5,
> >> > > > > >=10.0||<=-10 ];
> >> > > > > or
> >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> >> > > > >
> >> > > > > Thank you.
> >> > > > > Binyu
> >> > > > >
> >> > > > >
> >> > > >
> >> > > >
> >> > >
> >> > >
> >> >
> >> >
> >>
> >>
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Tue Aug 11 16:27:47 2020

Binyu,

I believe that WCOSS is down this week and unavailable... at least on
the
development side.

If you're able to grab that sample data and post it on the RAL
anonymous
ftp site, I could grab it from there for testing.

If that possible, please follow the instructions posted here:
http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp

Thanks,
John Halley Gotway

On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> Hello again,
>
> Following your suggestion, I tried to look for more sites for
verification
> by using prepbufr data. Unfortunately I always got "0 matching pair"
when I
> applied PB2NC and ensemble_stat. Here is the directory:
>
> cd
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
>
> % ./test_bufr.ecf >& bufr.log3
>
>
>  I tried to diagnose the problem by converting the prepbufr data to
grid
> data using point_to_grid, the script is:
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
>
> One of the output is:
>
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> regrid.prebufr.nc.txt
>
> It seems the PRES field does not look correct to me: it is over
~20000mb,
> which is impossible. I confirmed with prepbufr people, the unit
inside of
> the prepbufr is mb, and the magnitude should be <1000mb. But I can
not find
> where the error is. Any suggestions?
>
> Thank you.
> Binyu
>
>
>
> On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > To keep these issues straight, I just created a separate RT ticket
for
> > Tatiana to look at. Hopefully you received a copy of it.
> >
> > Thanks,
> > John
> >
> > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway
<johnhg at ucar.edu>
> > wrote:
> >
> > > Binyu,
> > >
> > > I'm guessing that the output is not actually the same. In the
VL1L2 and
> > > VCNT lines, the FCST_THRESH and OBS_THRESH columns should now
indicate
> > that
> > > you're using different thresholds: '>=1.0' should now be changed
to
> > > '<=-1.0||>=1.0'
> > >
> > > You should confirm that those have changed to be confident that
you
> > > actually ran the modified configuration. Once you've confirmed
that, if
> > the
> > > counts and stats are still the same, that's really just a
function of
> > your
> > > data... about which I have no idea.
> > >
> > > John
> > >
> > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > >>
> > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > >>
> > >> John,
> > >>
> > >> Should I send another ticket to "Tatiana" about my log error to
upload
> > >> point_stat*stat files to METViwer?
> > >>
> > >> I just tried the new config file (cat_thresh = [ <=-1.0||>=1.0,
> > >> <=-5.0||>=5.0 ];) and the output are the SAME as using the old
config
> > >> file (
> > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few
matching
> > >> pairs
> > >> (only one or two pairs)?
> > >>
> > >> Binyu
> > >>
> > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> > >> met_help at ucar.edu>
> > >> wrote:
> > >>
> > >> > Binyu,
> > >> >
> > >> > You're right, the cat_thresh for UGRD of >=1.0 would only
threshold
> > the
> > >> > positive values. If you want to apply both positive and
negative you
> > >> could
> > >> > use:
> > >> >
> > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > >> >
> > >> > Table 7.19 in the MET User's Guide describes the contents of
the
> VCNT
> > >> line
> > >> > type:
> > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > >> >
> > >> > Looking at the description of those statistics, it looks like
you
> may
> > be
> > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
DIR_ABSERR. And
> > >> > METviewer should be able to plot those statistics from the
VCNT line
> > >> type.
> > >> > That being said, I have never actually done this myself for a
model
> > >> > evaluation.
> > >> >
> > >> > If METviewer isn't loading VCNT lines or plotting the
statistics
> from
> > >> them,
> > >> > we'll need to let Tatiana Burek know so that she can fix it.
> > >> >
> > >> > Thanks,
> > >> > John
> > >> >
> > >> >
> > >> >
> > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT <
> > >> > met_help at ucar.edu> wrote:
> > >> >
> > >> > >
> > >> > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > >> > >
> > >> > > John,
> > >> > >
> > >> > > That makes sense for wind SPEED. Please look at the
point_stat
> > config:
> > >> > >
> > >> > >
> > >> > >
> > >> >
> > >>
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > >> > > PointStatConfig
> > >> > >
> > >> > > field = [
> > >> > >
> > >> > >      {
> > >> > >
> > >> > >         name       = "UGRD";
> > >> > >
> > >> > >         level      = [ "P250" ];
> > >> > >
> > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > >> > >
> > >> > >       },
> > >> > >
> > >> > >
> > >> > > the cat_thresh ONLY has positive too, does that mean both
> directions
> > >> will
> > >> > > be considered? (>= refers either ge 5 or le -5)?
> > >> > >
> > >> > >
> > >> > > One more question here: following your suggestion, I did
create
> > VL1L2
> > >> and
> > >> > > VCNT,but I got strange error (
> > >> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > >> when I
> > >> > > tried to load to METviewer ( I mentioned that in one of my
> previous
> > >> > email,
> > >> > > but didn't get reply). Is there any plot in METviwer that
can
> > compare
> > >> the
> > >> > > wind direction b/t obs. and forecast?
> > >> > >
> > >> > >
> > >> > > Thank you.
> > >> > >
> > >> > > Binyu
> > >> > >
> > >> > >
> > >> > >
> > >> > >
> > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT <
> > >> > > met_help at ucar.edu>
> > >> > > wrote:
> > >> > >
> > >> > > > Hi Binyu,
> > >> > > >
> > >> > > > The wind_thresh is a filtering threshold which is applied
to the
> > >> wind
> > >> > > > SPEED. While the individual U and V components can be
positive
> or
> > >> > > negative,
> > >> > > > the wind speed is always >= 0.
> > >> > > >
> > >> > > > The wind_thresh applies to the processing of the VL1L2,
VAL1L2,
> > and
> > >> > VCNT
> > >> > > > line types which process U/V pairs. Let's say you've set:
> > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > >> > > >
> > >> > > > Instead of seeing only 1 VL1L2 output line, you'll get
4... one
> > for
> > >> > each
> > >> > > > wind_thresh chosen.
> > >> > > > The first one (wind_thresh = NA) will include all U/V
pairs in
> the
> > >> > > masking
> > >> > > > area.
> > >> > > > The second one (wind_thresh ge1) will only include those
U/V
> pairs
> > >> > whose
> > >> > > > corresponding wind speed is at least 1 m/s.
> > >> > > > The third and fourth ones (wind_thresh ge5 and ge10) will
only
> > >> include
> > >> > > > those U/V pairs whose corresponding wind speed is at
least 5 m/s
> > or
> > >> > > 10m/s,
> > >> > > > respectively.
> > >> > > >
> > >> > > > The idea here is checking to see how the model performs
light
> and
> > >> > > variable
> > >> > > > winds (i.e. low wind speed) versus well-defined winds
(i.e. high
> > >> wind
> > >> > > > speed).
> > >> > > >
> > >> > > > Thanks,
> > >> > > > John
> > >> > > >
> > >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via
RT <
> > >> > > > met_help at ucar.edu> wrote:
> > >> > > >
> > >> > > > >
> > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted upon.
> > >> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > >> > > > >        Queue: met_help
> > >> > > > >      Subject: wind_thresh
> > >> > > > >        Owner: Nobody
> > >> > > > >   Requestors: binyu.wang at noaa.gov
> > >> > > > >       Status: new
> > >> > > > >  Ticket <URL:
> > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > >> > > >
> > >> > > > >
> > >> > > > >
> > >> > > > > Hello,
> > >> > > > >
> > >> > > > > I have a question about setting up wind_thresh in
config file:
> > >> > because
> > >> > > > the
> > >> > > > > wind can be negative and positive, so how to set them
> correctly?
> > >> > > > > e.g, which is correct:
> > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10
> > >> ||>=-10&&<=-5,
> > >> > > > > >=10.0||<=-10 ];
> > >> > > > > or
> > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > >> > > > >
> > >> > > > > Thank you.
> > >> > > > > Binyu
> > >> > > > >
> > >> > > > >
> > >> > > >
> > >> > > >
> > >> > >
> > >> > >
> > >> >
> > >> >
> > >>
> > >>
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Thu Aug 13 19:51:40 2020

Hello John,

There are some other files that may be involved to solve the problem,
so it
is better for me to just wait for WCOSS to come back...

Thank you,
Binyu

On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Binyu,
>
> I believe that WCOSS is down this week and unavailable... at least
on the
> development side.
>
> If you're able to grab that sample data and post it on the RAL
anonymous
> ftp site, I could grab it from there for testing.
>
> If that possible, please follow the instructions posted here:
>
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
>
> Thanks,
> John Halley Gotway
>
> On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> > Hello again,
> >
> > Following your suggestion, I tried to look for more sites for
> verification
> > by using prepbufr data. Unfortunately I always got "0 matching
pair"
> when I
> > applied PB2NC and ensemble_stat. Here is the directory:
> >
> > cd
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> >
> > % ./test_bufr.ecf >& bufr.log3
> >
> >
> >  I tried to diagnose the problem by converting the prepbufr data
to grid
> > data using point_to_grid, the script is:
> > /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> >
> > One of the output is:
> >
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > regrid.prebufr.nc.txt
> >
> > It seems the PRES field does not look correct to me: it is over
~20000mb,
> > which is impossible. I confirmed with prepbufr people, the unit
inside of
> > the prepbufr is mb, and the magnitude should be <1000mb. But I can
not
> find
> > where the error is. Any suggestions?
> >
> > Thank you.
> > Binyu
> >
> >
> >
> > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > To keep these issues straight, I just created a separate RT
ticket for
> > > Tatiana to look at. Hopefully you received a copy of it.
> > >
> > > Thanks,
> > > John
> > >
> > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway
<johnhg at ucar.edu>
> > > wrote:
> > >
> > > > Binyu,
> > > >
> > > > I'm guessing that the output is not actually the same. In the
VL1L2
> and
> > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns should now
> indicate
> > > that
> > > > you're using different thresholds: '>=1.0' should now be
changed to
> > > > '<=-1.0||>=1.0'
> > > >
> > > > You should confirm that those have changed to be confident
that you
> > > > actually ran the modified configuration. Once you've confirmed
that,
> if
> > > the
> > > > counts and stats are still the same, that's really just a
function of
> > > your
> > > > data... about which I have no idea.
> > > >
> > > > John
> > > >
> > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > >>
> > > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> > > >>
> > > >> John,
> > > >>
> > > >> Should I send another ticket to "Tatiana" about my log error
to
> upload
> > > >> point_stat*stat files to METViwer?
> > > >>
> > > >> I just tried the new config file (cat_thresh = [ <=-
1.0||>=1.0,
> > > >> <=-5.0||>=5.0 ];) and the output are the SAME as using the
old
> config
> > > >> file (
> > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few
> matching
> > > >> pairs
> > > >> (only one or two pairs)?
> > > >>
> > > >> Binyu
> > > >>
> > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> > > >> met_help at ucar.edu>
> > > >> wrote:
> > > >>
> > > >> > Binyu,
> > > >> >
> > > >> > You're right, the cat_thresh for UGRD of >=1.0 would only
> threshold
> > > the
> > > >> > positive values. If you want to apply both positive and
negative
> you
> > > >> could
> > > >> > use:
> > > >> >
> > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > >> >
> > > >> > Table 7.19 in the MET User's Guide describes the contents
of the
> > VCNT
> > > >> line
> > > >> > type:
> > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > >> >
> > > >> > Looking at the description of those statistics, it looks
like you
> > may
> > > be
> > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
DIR_ABSERR. And
> > > >> > METviewer should be able to plot those statistics from the
VCNT
> line
> > > >> type.
> > > >> > That being said, I have never actually done this myself for
a
> model
> > > >> > evaluation.
> > > >> >
> > > >> > If METviewer isn't loading VCNT lines or plotting the
statistics
> > from
> > > >> them,
> > > >> > we'll need to let Tatiana Burek know so that she can fix
it.
> > > >> >
> > > >> > Thanks,
> > > >> > John
> > > >> >
> > > >> >
> > > >> >
> > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT
<
> > > >> > met_help at ucar.edu> wrote:
> > > >> >
> > > >> > >
> > > >> > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > >> > >
> > > >> > > John,
> > > >> > >
> > > >> > > That makes sense for wind SPEED. Please look at the
point_stat
> > > config:
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> >
> > > >>
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > >> > > PointStatConfig
> > > >> > >
> > > >> > > field = [
> > > >> > >
> > > >> > >      {
> > > >> > >
> > > >> > >         name       = "UGRD";
> > > >> > >
> > > >> > >         level      = [ "P250" ];
> > > >> > >
> > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > >> > >
> > > >> > >       },
> > > >> > >
> > > >> > >
> > > >> > > the cat_thresh ONLY has positive too, does that mean both
> > directions
> > > >> will
> > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > >> > >
> > > >> > >
> > > >> > > One more question here: following your suggestion, I did
create
> > > VL1L2
> > > >> and
> > > >> > > VCNT,but I got strange error (
> > > >> > >
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > >> when I
> > > >> > > tried to load to METviewer ( I mentioned that in one of
my
> > previous
> > > >> > email,
> > > >> > > but didn't get reply). Is there any plot in METviwer that
can
> > > compare
> > > >> the
> > > >> > > wind direction b/t obs. and forecast?
> > > >> > >
> > > >> > >
> > > >> > > Thank you.
> > > >> > >
> > > >> > > Binyu
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT
<
> > > >> > > met_help at ucar.edu>
> > > >> > > wrote:
> > > >> > >
> > > >> > > > Hi Binyu,
> > > >> > > >
> > > >> > > > The wind_thresh is a filtering threshold which is
applied to
> the
> > > >> wind
> > > >> > > > SPEED. While the individual U and V components can be
positive
> > or
> > > >> > > negative,
> > > >> > > > the wind speed is always >= 0.
> > > >> > > >
> > > >> > > > The wind_thresh applies to the processing of the VL1L2,
> VAL1L2,
> > > and
> > > >> > VCNT
> > > >> > > > line types which process U/V pairs. Let's say you've
set:
> > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > >> > > >
> > > >> > > > Instead of seeing only 1 VL1L2 output line, you'll get
4...
> one
> > > for
> > > >> > each
> > > >> > > > wind_thresh chosen.
> > > >> > > > The first one (wind_thresh = NA) will include all U/V
pairs in
> > the
> > > >> > > masking
> > > >> > > > area.
> > > >> > > > The second one (wind_thresh ge1) will only include
those U/V
> > pairs
> > > >> > whose
> > > >> > > > corresponding wind speed is at least 1 m/s.
> > > >> > > > The third and fourth ones (wind_thresh ge5 and ge10)
will only
> > > >> include
> > > >> > > > those U/V pairs whose corresponding wind speed is at
least 5
> m/s
> > > or
> > > >> > > 10m/s,
> > > >> > > > respectively.
> > > >> > > >
> > > >> > > > The idea here is checking to see how the model performs
light
> > and
> > > >> > > variable
> > > >> > > > winds (i.e. low wind speed) versus well-defined winds
(i.e.
> high
> > > >> wind
> > > >> > > > speed).
> > > >> > > >
> > > >> > > > Thanks,
> > > >> > > > John
> > > >> > > >
> > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via
RT <
> > > >> > > > met_help at ucar.edu> wrote:
> > > >> > > >
> > > >> > > > >
> > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted
upon.
> > > >> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > > >> > > > >        Queue: met_help
> > > >> > > > >      Subject: wind_thresh
> > > >> > > > >        Owner: Nobody
> > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > >> > > > >       Status: new
> > > >> > > > >  Ticket <URL:
> > > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > >> > > >
> > > >> > > > >
> > > >> > > > >
> > > >> > > > > Hello,
> > > >> > > > >
> > > >> > > > > I have a question about setting up wind_thresh in
config
> file:
> > > >> > because
> > > >> > > > the
> > > >> > > > > wind can be negative and positive, so how to set them
> > correctly?
> > > >> > > > > e.g, which is correct:
> > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10
> > > >> ||>=-10&&<=-5,
> > > >> > > > > >=10.0||<=-10 ];
> > > >> > > > > or
> > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > > >> > > > >
> > > >> > > > > Thank you.
> > > >> > > > > Binyu
> > > >> > > > >
> > > >> > > > >
> > > >> > > >
> > > >> > > >
> > > >> > >
> > > >> > >
> > > >> >
> > > >> >
> > > >>
> > > >>
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Mon Aug 17 15:41:03 2020

OK, please let me know when there's something you'd like me to look
at.

John

On Thu, Aug 13, 2020 at 7:52 PM binyu.wang at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> Hello John,
>
> There are some other files that may be involved to solve the
problem, so it
> is better for me to just wait for WCOSS to come back...
>
> Thank you,
> Binyu
>
> On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Binyu,
> >
> > I believe that WCOSS is down this week and unavailable... at least
on the
> > development side.
> >
> > If you're able to grab that sample data and post it on the RAL
anonymous
> > ftp site, I could grab it from there for testing.
> >
> > If that possible, please follow the instructions posted here:
> >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> >
> > Thanks,
> > John Halley Gotway
> >
> > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > >
> > > Hello again,
> > >
> > > Following your suggestion, I tried to look for more sites for
> > verification
> > > by using prepbufr data. Unfortunately I always got "0 matching
pair"
> > when I
> > > applied PB2NC and ensemble_stat. Here is the directory:
> > >
> > > cd
> > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > >
> > > % ./test_bufr.ecf >& bufr.log3
> > >
> > >
> > >  I tried to diagnose the problem by converting the prepbufr data
to
> grid
> > > data using point_to_grid, the script is:
> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > >
> > > One of the output is:
> > >
> > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > regrid.prebufr.nc.txt
> > >
> > > It seems the PRES field does not look correct to me: it is over
> ~20000mb,
> > > which is impossible. I confirmed with prepbufr people, the unit
inside
> of
> > > the prepbufr is mb, and the magnitude should be <1000mb. But I
can not
> > find
> > > where the error is. Any suggestions?
> > >
> > > Thank you.
> > > Binyu
> > >
> > >
> > >
> > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > > To keep these issues straight, I just created a separate RT
ticket
> for
> > > > Tatiana to look at. Hopefully you received a copy of it.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway
<johnhg at ucar.edu>
> > > > wrote:
> > > >
> > > > > Binyu,
> > > > >
> > > > > I'm guessing that the output is not actually the same. In
the VL1L2
> > and
> > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns should
now
> > indicate
> > > > that
> > > > > you're using different thresholds: '>=1.0' should now be
changed to
> > > > > '<=-1.0||>=1.0'
> > > > >
> > > > > You should confirm that those have changed to be confident
that you
> > > > > actually ran the modified configuration. Once you've
confirmed
> that,
> > if
> > > > the
> > > > > counts and stats are still the same, that's really just a
function
> of
> > > > your
> > > > > data... about which I have no idea.
> > > > >
> > > > > John
> > > > >
> > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > >>
> > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > >>
> > > > >> John,
> > > > >>
> > > > >> Should I send another ticket to "Tatiana" about my log
error to
> > upload
> > > > >> point_stat*stat files to METViwer?
> > > > >>
> > > > >> I just tried the new config file (cat_thresh = [ <=-
1.0||>=1.0,
> > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as using the
old
> > config
> > > > >> file (
> > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too
few
> > matching
> > > > >> pairs
> > > > >> (only one or two pairs)?
> > > > >>
> > > > >> Binyu
> > > > >>
> > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> > > > >> met_help at ucar.edu>
> > > > >> wrote:
> > > > >>
> > > > >> > Binyu,
> > > > >> >
> > > > >> > You're right, the cat_thresh for UGRD of >=1.0 would only
> > threshold
> > > > the
> > > > >> > positive values. If you want to apply both positive and
negative
> > you
> > > > >> could
> > > > >> > use:
> > > > >> >
> > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > >> >
> > > > >> > Table 7.19 in the MET User's Guide describes the contents
of the
> > > VCNT
> > > > >> line
> > > > >> > type:
> > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > >> >
> > > > >> > Looking at the description of those statistics, it looks
like
> you
> > > may
> > > > be
> > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
DIR_ABSERR.
> And
> > > > >> > METviewer should be able to plot those statistics from
the VCNT
> > line
> > > > >> type.
> > > > >> > That being said, I have never actually done this myself
for a
> > model
> > > > >> > evaluation.
> > > > >> >
> > > > >> > If METviewer isn't loading VCNT lines or plotting the
statistics
> > > from
> > > > >> them,
> > > > >> > we'll need to let Tatiana Burek know so that she can fix
it.
> > > > >> >
> > > > >> > Thanks,
> > > > >> > John
> > > > >> >
> > > > >> >
> > > > >> >
> > > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via
RT <
> > > > >> > met_help at ucar.edu> wrote:
> > > > >> >
> > > > >> > >
> > > > >> > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >
> > > > >> > >
> > > > >> > > John,
> > > > >> > >
> > > > >> > > That makes sense for wind SPEED. Please look at the
point_stat
> > > > config:
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> >
> > > > >>
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > >> > > PointStatConfig
> > > > >> > >
> > > > >> > > field = [
> > > > >> > >
> > > > >> > >      {
> > > > >> > >
> > > > >> > >         name       = "UGRD";
> > > > >> > >
> > > > >> > >         level      = [ "P250" ];
> > > > >> > >
> > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > >> > >
> > > > >> > >       },
> > > > >> > >
> > > > >> > >
> > > > >> > > the cat_thresh ONLY has positive too, does that mean
both
> > > directions
> > > > >> will
> > > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > > >> > >
> > > > >> > >
> > > > >> > > One more question here: following your suggestion, I
did
> create
> > > > VL1L2
> > > > >> and
> > > > >> > > VCNT,but I got strange error (
> > > > >> > >
> > /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > >> when I
> > > > >> > > tried to load to METviewer ( I mentioned that in one of
my
> > > previous
> > > > >> > email,
> > > > >> > > but didn't get reply). Is there any plot in METviwer
that can
> > > > compare
> > > > >> the
> > > > >> > > wind direction b/t obs. and forecast?
> > > > >> > >
> > > > >> > >
> > > > >> > > Thank you.
> > > > >> > >
> > > > >> > > Binyu
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via
RT <
> > > > >> > > met_help at ucar.edu>
> > > > >> > > wrote:
> > > > >> > >
> > > > >> > > > Hi Binyu,
> > > > >> > > >
> > > > >> > > > The wind_thresh is a filtering threshold which is
applied to
> > the
> > > > >> wind
> > > > >> > > > SPEED. While the individual U and V components can be
> positive
> > > or
> > > > >> > > negative,
> > > > >> > > > the wind speed is always >= 0.
> > > > >> > > >
> > > > >> > > > The wind_thresh applies to the processing of the
VL1L2,
> > VAL1L2,
> > > > and
> > > > >> > VCNT
> > > > >> > > > line types which process U/V pairs. Let's say you've
set:
> > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > >> > > >
> > > > >> > > > Instead of seeing only 1 VL1L2 output line, you'll
get 4...
> > one
> > > > for
> > > > >> > each
> > > > >> > > > wind_thresh chosen.
> > > > >> > > > The first one (wind_thresh = NA) will include all U/V
pairs
> in
> > > the
> > > > >> > > masking
> > > > >> > > > area.
> > > > >> > > > The second one (wind_thresh ge1) will only include
those U/V
> > > pairs
> > > > >> > whose
> > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > >> > > > The third and fourth ones (wind_thresh ge5 and ge10)
will
> only
> > > > >> include
> > > > >> > > > those U/V pairs whose corresponding wind speed is at
least 5
> > m/s
> > > > or
> > > > >> > > 10m/s,
> > > > >> > > > respectively.
> > > > >> > > >
> > > > >> > > > The idea here is checking to see how the model
performs
> light
> > > and
> > > > >> > > variable
> > > > >> > > > winds (i.e. low wind speed) versus well-defined winds
(i.e.
> > high
> > > > >> wind
> > > > >> > > > speed).
> > > > >> > > >
> > > > >> > > > Thanks,
> > > > >> > > > John
> > > > >> > > >
> > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov
via RT
> <
> > > > >> > > > met_help at ucar.edu> wrote:
> > > > >> > > >
> > > > >> > > > >
> > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted
upon.
> > > > >> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > > > >> > > > >        Queue: met_help
> > > > >> > > > >      Subject: wind_thresh
> > > > >> > > > >        Owner: Nobody
> > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > >> > > > >       Status: new
> > > > >> > > > >  Ticket <URL:
> > > > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > >> > > >
> > > > >> > > > >
> > > > >> > > > >
> > > > >> > > > > Hello,
> > > > >> > > > >
> > > > >> > > > > I have a question about setting up wind_thresh in
config
> > file:
> > > > >> > because
> > > > >> > > > the
> > > > >> > > > > wind can be negative and positive, so how to set
them
> > > correctly?
> > > > >> > > > > e.g, which is correct:
> > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
>=5.0&&<=10
> > > > >> ||>=-10&&<=-5,
> > > > >> > > > > >=10.0||<=-10 ];
> > > > >> > > > > or
> > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > > > >> > > > >
> > > > >> > > > > Thank you.
> > > > >> > > > > Binyu
> > > > >> > > > >
> > > > >> > > > >
> > > > >> > > >
> > > > >> > > >
> > > > >> > >
> > > > >> > >
> > > > >> >
> > > > >> >
> > > > >>
> > > > >>
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Mon Aug 17 16:50:05 2020

I just uploaded some files for this question under the same directory
"wang". I do the point2grid with the prepbufr data because I got " 0
pairing" using "ensemble_stat".

I converted the prepbufr data to grid data using point_to_grid,
the script is "pt2grid.sh"
the model file: gec00.t00z.pgrb2p5.0p50.f000
the prepbufr file:prepbufr.gdas.2020020100

config file: PB2NCConfig_gens
one output I got: regrid.prebufr.nc.txt

The PRES field does not look correct to me: it is over ~20000mb, which
is
impossible. I can not figure it out why?

Binyu

>
> some other files that may be involved to solve the problem, so it
> > is better for me to just wait for WCOSS to come back...
> >
> > Thank you,
> > Binyu
> >
> > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Binyu,
> > >
> > > I believe that WCOSS is down this week and unavailable... at
least on
> the
> > > development side.
> > >
> > > If you're able to grab that sample data and post it on the RAL
> anonymous
> > > ftp site, I could grab it from there for testing.
> > >
> > > If that possible, please follow the instructions posted here:
> > >
> > >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> > >
> > > Thanks,
> > > John Halley Gotway
> > >
> > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> > > >
> > > > Hello again,
> > > >
> > > > Following your suggestion, I tried to look for more sites for
> > > verification
> > > > by using prepbufr data. Unfortunately I always got "0 matching
pair"
> > > when I
> > > > applied PB2NC and ensemble_stat. Here is the directory:
> > > >
> > > > cd
> > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > > >
> > > > % ./test_bufr.ecf >& bufr.log3
> > > >
> > > >
> > > >  I tried to diagnose the problem by converting the prepbufr
data to
> > grid
> > > > data using point_to_grid, the script is:
> > > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > > >
> > > > One of the output is:
> > > >
> > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > > regrid.prebufr.nc.txt
> > > >
> > > > It seems the PRES field does not look correct to me: it is
over
> > ~20000mb,
> > > > which is impossible. I confirmed with prepbufr people, the
unit
> inside
> > of
> > > > the prepbufr is mb, and the magnitude should be <1000mb. But I
can
> not
> > > find
> > > > where the error is. Any suggestions?
> > > >
> > > > Thank you.
> > > > Binyu
> > > >
> > > >
> > > >
> > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > > To keep these issues straight, I just created a separate RT
ticket
> > for
> > > > > Tatiana to look at. Hopefully you received a copy of it.
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <
> johnhg at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > > Binyu,
> > > > > >
> > > > > > I'm guessing that the output is not actually the same. In
the
> VL1L2
> > > and
> > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns should
now
> > > indicate
> > > > > that
> > > > > > you're using different thresholds: '>=1.0' should now be
changed
> to
> > > > > > '<=-1.0||>=1.0'
> > > > > >
> > > > > > You should confirm that those have changed to be confident
that
> you
> > > > > > actually ran the modified configuration. Once you've
confirmed
> > that,
> > > if
> > > > > the
> > > > > > counts and stats are still the same, that's really just a
> function
> > of
> > > > > your
> > > > > > data... about which I have no idea.
> > > > > >
> > > > > > John
> > > > > >
> > > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT
<
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > >>
> > > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > > >>
> > > > > >> John,
> > > > > >>
> > > > > >> Should I send another ticket to "Tatiana" about my log
error to
> > > upload
> > > > > >> point_stat*stat files to METViwer?
> > > > > >>
> > > > > >> I just tried the new config file (cat_thresh = [ <=-
1.0||>=1.0,
> > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as using
the old
> > > config
> > > > > >> file (
> > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too
few
> > > matching
> > > > > >> pairs
> > > > > >> (only one or two pairs)?
> > > > > >>
> > > > > >> Binyu
> > > > > >>
> > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT
<
> > > > > >> met_help at ucar.edu>
> > > > > >> wrote:
> > > > > >>
> > > > > >> > Binyu,
> > > > > >> >
> > > > > >> > You're right, the cat_thresh for UGRD of >=1.0 would
only
> > > threshold
> > > > > the
> > > > > >> > positive values. If you want to apply both positive and
> negative
> > > you
> > > > > >> could
> > > > > >> > use:
> > > > > >> >
> > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > > >> >
> > > > > >> > Table 7.19 in the MET User's Guide describes the
contents of
> the
> > > > VCNT
> > > > > >> line
> > > > > >> > type:
> > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > > >> >
> > > > > >> > Looking at the description of those statistics, it
looks like
> > you
> > > > may
> > > > > be
> > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
DIR_ABSERR.
> > And
> > > > > >> > METviewer should be able to plot those statistics from
the
> VCNT
> > > line
> > > > > >> type.
> > > > > >> > That being said, I have never actually done this myself
for a
> > > model
> > > > > >> > evaluation.
> > > > > >> >
> > > > > >> > If METviewer isn't loading VCNT lines or plotting the
> statistics
> > > > from
> > > > > >> them,
> > > > > >> > we'll need to let Tatiana Burek know so that she can
fix it.
> > > > > >> >
> > > > > >> > Thanks,
> > > > > >> > John
> > > > > >> >
> > > > > >> >
> > > > > >> >
> > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via
RT <
> > > > > >> > met_help at ucar.edu> wrote:
> > > > > >> >
> > > > > >> > >
> > > > > >> > > <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > >
> > > > > >> > >
> > > > > >> > > John,
> > > > > >> > >
> > > > > >> > > That makes sense for wind SPEED. Please look at the
> point_stat
> > > > > config:
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> >
> > > > > >>
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > > >> > > PointStatConfig
> > > > > >> > >
> > > > > >> > > field = [
> > > > > >> > >
> > > > > >> > >      {
> > > > > >> > >
> > > > > >> > >         name       = "UGRD";
> > > > > >> > >
> > > > > >> > >         level      = [ "P250" ];
> > > > > >> > >
> > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > > >> > >
> > > > > >> > >       },
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > the cat_thresh ONLY has positive too, does that mean
both
> > > > directions
> > > > > >> will
> > > > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > One more question here: following your suggestion, I
did
> > create
> > > > > VL1L2
> > > > > >> and
> > > > > >> > > VCNT,but I got strange error (
> > > > > >> > >
> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > > >> when I
> > > > > >> > > tried to load to METviewer ( I mentioned that in one
of my
> > > > previous
> > > > > >> > email,
> > > > > >> > > but didn't get reply). Is there any plot in METviwer
that
> can
> > > > > compare
> > > > > >> the
> > > > > >> > > wind direction b/t obs. and forecast?
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > Thank you.
> > > > > >> > >
> > > > > >> > > Binyu
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway
via RT <
> > > > > >> > > met_help at ucar.edu>
> > > > > >> > > wrote:
> > > > > >> > >
> > > > > >> > > > Hi Binyu,
> > > > > >> > > >
> > > > > >> > > > The wind_thresh is a filtering threshold which is
applied
> to
> > > the
> > > > > >> wind
> > > > > >> > > > SPEED. While the individual U and V components can
be
> > positive
> > > > or
> > > > > >> > > negative,
> > > > > >> > > > the wind speed is always >= 0.
> > > > > >> > > >
> > > > > >> > > > The wind_thresh applies to the processing of the
VL1L2,
> > > VAL1L2,
> > > > > and
> > > > > >> > VCNT
> > > > > >> > > > line types which process U/V pairs. Let's say
you've set:
> > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > > >> > > >
> > > > > >> > > > Instead of seeing only 1 VL1L2 output line, you'll
get
> 4...
> > > one
> > > > > for
> > > > > >> > each
> > > > > >> > > > wind_thresh chosen.
> > > > > >> > > > The first one (wind_thresh = NA) will include all
U/V
> pairs
> > in
> > > > the
> > > > > >> > > masking
> > > > > >> > > > area.
> > > > > >> > > > The second one (wind_thresh ge1) will only include
those
> U/V
> > > > pairs
> > > > > >> > whose
> > > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > > >> > > > The third and fourth ones (wind_thresh ge5 and
ge10) will
> > only
> > > > > >> include
> > > > > >> > > > those U/V pairs whose corresponding wind speed is
at
> least 5
> > > m/s
> > > > > or
> > > > > >> > > 10m/s,
> > > > > >> > > > respectively.
> > > > > >> > > >
> > > > > >> > > > The idea here is checking to see how the model
performs
> > light
> > > > and
> > > > > >> > > variable
> > > > > >> > > > winds (i.e. low wind speed) versus well-defined
winds
> (i.e.
> > > high
> > > > > >> wind
> > > > > >> > > > speed).
> > > > > >> > > >
> > > > > >> > > > Thanks,
> > > > > >> > > > John
> > > > > >> > > >
> > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov
via
> RT
> > <
> > > > > >> > > > met_help at ucar.edu> wrote:
> > > > > >> > > >
> > > > > >> > > > >
> > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted
upon.
> > > > > >> > > > > Transaction: Ticket created by
binyu.wang at noaa.gov
> > > > > >> > > > >        Queue: met_help
> > > > > >> > > > >      Subject: wind_thresh
> > > > > >> > > > >        Owner: Nobody
> > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > > >> > > > >       Status: new
> > > > > >> > > > >  Ticket <URL:
> > > > > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > > >> > > >
> > > > > >> > > > >
> > > > > >> > > > >
> > > > > >> > > > > Hello,
> > > > > >> > > > >
> > > > > >> > > > > I have a question about setting up wind_thresh in
config
> > > file:
> > > > > >> > because
> > > > > >> > > > the
> > > > > >> > > > > wind can be negative and positive, so how to set
them
> > > > correctly?
> > > > > >> > > > > e.g, which is correct:
> > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
>=5.0&&<=10
> > > > > >> ||>=-10&&<=-5,
> > > > > >> > > > > >=10.0||<=-10 ];
> > > > > >> > > > > or
> > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20
];
> > > > > >> > > > >
> > > > > >> > > > > Thank you.
> > > > > >> > > > > Binyu
> > > > > >> > > > >
> > > > > >> > > > >
> > > > > >> > > >
> > > > > >> > > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> >
> > > > > >> >
> > > > > >>
> > > > > >>
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Thu Aug 27 14:28:13 2020

Hello John,

Just want to remind that I am still waiting for the answer. I am sorry
to
keep asking questions: I guess I am the one with most of the
questions.

On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Binyu,
>
> I believe that WCOSS is down this week and unavailable... at least
on the
> development side.
>
> If you're able to grab that sample data and post it on the RAL
anonymous
> ftp site, I could grab it from there for testing.
>
> If that possible, please follow the instructions posted here:
>
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
>
> Thanks,
> John Halley Gotway
>
> On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> > Hello again,
> >
> > Following your suggestion, I tried to look for more sites for
> verification
> > by using prepbufr data. Unfortunately I always got "0 matching
pair"
> when I
> > applied PB2NC and ensemble_stat. Here is the directory:
> >
> > cd
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> >
> > % ./test_bufr.ecf >& bufr.log3
> >
> >
> >  I tried to diagnose the problem by converting the prepbufr data
to grid
> > data using point_to_grid, the script is:
> > /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> >
> > One of the output is:
> >
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > regrid.prebufr.nc.txt
> >
> > It seems the PRES field does not look correct to me: it is over
~20000mb,
> > which is impossible. I confirmed with prepbufr people, the unit
inside of
> > the prepbufr is mb, and the magnitude should be <1000mb. But I can
not
> find
> > where the error is. Any suggestions?
> >
> > Thank you.
> > Binyu
> >
> >
> >
> > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > To keep these issues straight, I just created a separate RT
ticket for
> > > Tatiana to look at. Hopefully you received a copy of it.
> > >
> > > Thanks,
> > > John
> > >
> > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway
<johnhg at ucar.edu>
> > > wrote:
> > >
> > > > Binyu,
> > > >
> > > > I'm guessing that the output is not actually the same. In the
VL1L2
> and
> > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns should now
> indicate
> > > that
> > > > you're using different thresholds: '>=1.0' should now be
changed to
> > > > '<=-1.0||>=1.0'
> > > >
> > > > You should confirm that those have changed to be confident
that you
> > > > actually ran the modified configuration. Once you've confirmed
that,
> if
> > > the
> > > > counts and stats are still the same, that's really just a
function of
> > > your
> > > > data... about which I have no idea.
> > > >
> > > > John
> > > >
> > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > >>
> > > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> > > >>
> > > >> John,
> > > >>
> > > >> Should I send another ticket to "Tatiana" about my log error
to
> upload
> > > >> point_stat*stat files to METViwer?
> > > >>
> > > >> I just tried the new config file (cat_thresh = [ <=-
1.0||>=1.0,
> > > >> <=-5.0||>=5.0 ];) and the output are the SAME as using the
old
> config
> > > >> file (
> > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too few
> matching
> > > >> pairs
> > > >> (only one or two pairs)?
> > > >>
> > > >> Binyu
> > > >>
> > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> > > >> met_help at ucar.edu>
> > > >> wrote:
> > > >>
> > > >> > Binyu,
> > > >> >
> > > >> > You're right, the cat_thresh for UGRD of >=1.0 would only
> threshold
> > > the
> > > >> > positive values. If you want to apply both positive and
negative
> you
> > > >> could
> > > >> > use:
> > > >> >
> > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > >> >
> > > >> > Table 7.19 in the MET User's Guide describes the contents
of the
> > VCNT
> > > >> line
> > > >> > type:
> > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > >> >
> > > >> > Looking at the description of those statistics, it looks
like you
> > may
> > > be
> > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
DIR_ABSERR. And
> > > >> > METviewer should be able to plot those statistics from the
VCNT
> line
> > > >> type.
> > > >> > That being said, I have never actually done this myself for
a
> model
> > > >> > evaluation.
> > > >> >
> > > >> > If METviewer isn't loading VCNT lines or plotting the
statistics
> > from
> > > >> them,
> > > >> > we'll need to let Tatiana Burek know so that she can fix
it.
> > > >> >
> > > >> > Thanks,
> > > >> > John
> > > >> >
> > > >> >
> > > >> >
> > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via RT
<
> > > >> > met_help at ucar.edu> wrote:
> > > >> >
> > > >> > >
> > > >> > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > >> > >
> > > >> > > John,
> > > >> > >
> > > >> > > That makes sense for wind SPEED. Please look at the
point_stat
> > > config:
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> >
> > > >>
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > >> > > PointStatConfig
> > > >> > >
> > > >> > > field = [
> > > >> > >
> > > >> > >      {
> > > >> > >
> > > >> > >         name       = "UGRD";
> > > >> > >
> > > >> > >         level      = [ "P250" ];
> > > >> > >
> > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > >> > >
> > > >> > >       },
> > > >> > >
> > > >> > >
> > > >> > > the cat_thresh ONLY has positive too, does that mean both
> > directions
> > > >> will
> > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > >> > >
> > > >> > >
> > > >> > > One more question here: following your suggestion, I did
create
> > > VL1L2
> > > >> and
> > > >> > > VCNT,but I got strange error (
> > > >> > >
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > >> when I
> > > >> > > tried to load to METviewer ( I mentioned that in one of
my
> > previous
> > > >> > email,
> > > >> > > but didn't get reply). Is there any plot in METviwer that
can
> > > compare
> > > >> the
> > > >> > > wind direction b/t obs. and forecast?
> > > >> > >
> > > >> > >
> > > >> > > Thank you.
> > > >> > >
> > > >> > > Binyu
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> > >
> > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via RT
<
> > > >> > > met_help at ucar.edu>
> > > >> > > wrote:
> > > >> > >
> > > >> > > > Hi Binyu,
> > > >> > > >
> > > >> > > > The wind_thresh is a filtering threshold which is
applied to
> the
> > > >> wind
> > > >> > > > SPEED. While the individual U and V components can be
positive
> > or
> > > >> > > negative,
> > > >> > > > the wind speed is always >= 0.
> > > >> > > >
> > > >> > > > The wind_thresh applies to the processing of the VL1L2,
> VAL1L2,
> > > and
> > > >> > VCNT
> > > >> > > > line types which process U/V pairs. Let's say you've
set:
> > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > >> > > >
> > > >> > > > Instead of seeing only 1 VL1L2 output line, you'll get
4...
> one
> > > for
> > > >> > each
> > > >> > > > wind_thresh chosen.
> > > >> > > > The first one (wind_thresh = NA) will include all U/V
pairs in
> > the
> > > >> > > masking
> > > >> > > > area.
> > > >> > > > The second one (wind_thresh ge1) will only include
those U/V
> > pairs
> > > >> > whose
> > > >> > > > corresponding wind speed is at least 1 m/s.
> > > >> > > > The third and fourth ones (wind_thresh ge5 and ge10)
will only
> > > >> include
> > > >> > > > those U/V pairs whose corresponding wind speed is at
least 5
> m/s
> > > or
> > > >> > > 10m/s,
> > > >> > > > respectively.
> > > >> > > >
> > > >> > > > The idea here is checking to see how the model performs
light
> > and
> > > >> > > variable
> > > >> > > > winds (i.e. low wind speed) versus well-defined winds
(i.e.
> high
> > > >> wind
> > > >> > > > speed).
> > > >> > > >
> > > >> > > > Thanks,
> > > >> > > > John
> > > >> > > >
> > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov via
RT <
> > > >> > > > met_help at ucar.edu> wrote:
> > > >> > > >
> > > >> > > > >
> > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted
upon.
> > > >> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > > >> > > > >        Queue: met_help
> > > >> > > > >      Subject: wind_thresh
> > > >> > > > >        Owner: Nobody
> > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > >> > > > >       Status: new
> > > >> > > > >  Ticket <URL:
> > > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > >> > > >
> > > >> > > > >
> > > >> > > > >
> > > >> > > > > Hello,
> > > >> > > > >
> > > >> > > > > I have a question about setting up wind_thresh in
config
> file:
> > > >> > because
> > > >> > > > the
> > > >> > > > > wind can be negative and positive, so how to set them
> > correctly?
> > > >> > > > > e.g, which is correct:
> > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,  >=5.0&&<=10
> > > >> ||>=-10&&<=-5,
> > > >> > > > > >=10.0||<=-10 ];
> > > >> > > > > or
> > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > > >> > > > >
> > > >> > > > > Thank you.
> > > >> > > > > Binyu
> > > >> > > > >
> > > >> > > > >
> > > >> > > >
> > > >> > > >
> > > >> > >
> > > >> > >
> > > >> >
> > > >> >
> > > >>
> > > >>
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Thu Aug 27 15:00:28 2020

Binyu,

Yes, I was just writing up this response...

OK, I logged onto WCOSS and was able to replicate the behavior you
describe.

Here's what I found...

met-9.1_beta2 does run with this data but met-9.1 does not. But that's
because there is a bug in met-9.1_beta2 that was fixed in met-9.1.

Take a look at this line in point_stat.cc:
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861

The bug in beta2 caused this logic not be enforced because "is_vgrd"
never
evaluated to true.

But the real point here is that by convention, MET expects that when
processing winds as vectors, each point observation of UGRD is
immediately
followed by a VGRD observation at the same level. We enforce this
requirement to avoid the need of wasting processing time searching
through
all the U/V pairs to match them up. In the data you passed to
point_stat
your have 175 observations of UGRD followed by 175 observations of
VGRD. It
just so happens that the order of the levels in both is identical...
but
met-9.1_beta2 didn't actually check that to be sure! So if you made a
mistake in that regard, met-9.1_beta2 would not have caught it.

To test things out, I reordered your data to interleave UGRD and VGRD
for
each level. And the resulting file ran fine through point_stat. I've
attached that NetCDF file here.

Hope that helps clarify.

Thanks,
John

On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> Hello John,
>
> Just want to remind that I am still waiting for the answer. I am
sorry to
> keep asking questions: I guess I am the one with most of the
questions.
>
>
>
> On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Binyu,
> >
> > I believe that WCOSS is down this week and unavailable... at least
on the
> > development side.
> >
> > If you're able to grab that sample data and post it on the RAL
anonymous
> > ftp site, I could grab it from there for testing.
> >
> > If that possible, please follow the instructions posted here:
> >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> >
> > Thanks,
> > John Halley Gotway
> >
> > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > >
> > > Hello again,
> > >
> > > Following your suggestion, I tried to look for more sites for
> > verification
> > > by using prepbufr data. Unfortunately I always got "0 matching
pair"
> > when I
> > > applied PB2NC and ensemble_stat. Here is the directory:
> > >
> > > cd
> > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > >
> > > % ./test_bufr.ecf >& bufr.log3
> > >
> > >
> > >  I tried to diagnose the problem by converting the prepbufr data
to
> grid
> > > data using point_to_grid, the script is:
> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > >
> > > One of the output is:
> > >
> > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > regrid.prebufr.nc.txt
> > >
> > > It seems the PRES field does not look correct to me: it is over
> ~20000mb,
> > > which is impossible. I confirmed with prepbufr people, the unit
inside
> of
> > > the prepbufr is mb, and the magnitude should be <1000mb. But I
can not
> > find
> > > where the error is. Any suggestions?
> > >
> > > Thank you.
> > > Binyu
> > >
> > >
> > >
> > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > > To keep these issues straight, I just created a separate RT
ticket
> for
> > > > Tatiana to look at. Hopefully you received a copy of it.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway
<johnhg at ucar.edu>
> > > > wrote:
> > > >
> > > > > Binyu,
> > > > >
> > > > > I'm guessing that the output is not actually the same. In
the VL1L2
> > and
> > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns should
now
> > indicate
> > > > that
> > > > > you're using different thresholds: '>=1.0' should now be
changed to
> > > > > '<=-1.0||>=1.0'
> > > > >
> > > > > You should confirm that those have changed to be confident
that you
> > > > > actually ran the modified configuration. Once you've
confirmed
> that,
> > if
> > > > the
> > > > > counts and stats are still the same, that's really just a
function
> of
> > > > your
> > > > > data... about which I have no idea.
> > > > >
> > > > > John
> > > > >
> > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > >>
> > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > >>
> > > > >> John,
> > > > >>
> > > > >> Should I send another ticket to "Tatiana" about my log
error to
> > upload
> > > > >> point_stat*stat files to METViwer?
> > > > >>
> > > > >> I just tried the new config file (cat_thresh = [ <=-
1.0||>=1.0,
> > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as using the
old
> > config
> > > > >> file (
> > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too
few
> > matching
> > > > >> pairs
> > > > >> (only one or two pairs)?
> > > > >>
> > > > >> Binyu
> > > > >>
> > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT <
> > > > >> met_help at ucar.edu>
> > > > >> wrote:
> > > > >>
> > > > >> > Binyu,
> > > > >> >
> > > > >> > You're right, the cat_thresh for UGRD of >=1.0 would only
> > threshold
> > > > the
> > > > >> > positive values. If you want to apply both positive and
negative
> > you
> > > > >> could
> > > > >> > use:
> > > > >> >
> > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > >> >
> > > > >> > Table 7.19 in the MET User's Guide describes the contents
of the
> > > VCNT
> > > > >> line
> > > > >> > type:
> > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > >> >
> > > > >> > Looking at the description of those statistics, it looks
like
> you
> > > may
> > > > be
> > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
DIR_ABSERR.
> And
> > > > >> > METviewer should be able to plot those statistics from
the VCNT
> > line
> > > > >> type.
> > > > >> > That being said, I have never actually done this myself
for a
> > model
> > > > >> > evaluation.
> > > > >> >
> > > > >> > If METviewer isn't loading VCNT lines or plotting the
statistics
> > > from
> > > > >> them,
> > > > >> > we'll need to let Tatiana Burek know so that she can fix
it.
> > > > >> >
> > > > >> > Thanks,
> > > > >> > John
> > > > >> >
> > > > >> >
> > > > >> >
> > > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via
RT <
> > > > >> > met_help at ucar.edu> wrote:
> > > > >> >
> > > > >> > >
> > > > >> > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >
> > > > >> > >
> > > > >> > > John,
> > > > >> > >
> > > > >> > > That makes sense for wind SPEED. Please look at the
point_stat
> > > > config:
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> >
> > > > >>
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > >> > > PointStatConfig
> > > > >> > >
> > > > >> > > field = [
> > > > >> > >
> > > > >> > >      {
> > > > >> > >
> > > > >> > >         name       = "UGRD";
> > > > >> > >
> > > > >> > >         level      = [ "P250" ];
> > > > >> > >
> > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > >> > >
> > > > >> > >       },
> > > > >> > >
> > > > >> > >
> > > > >> > > the cat_thresh ONLY has positive too, does that mean
both
> > > directions
> > > > >> will
> > > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > > >> > >
> > > > >> > >
> > > > >> > > One more question here: following your suggestion, I
did
> create
> > > > VL1L2
> > > > >> and
> > > > >> > > VCNT,but I got strange error (
> > > > >> > >
> > /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > >> when I
> > > > >> > > tried to load to METviewer ( I mentioned that in one of
my
> > > previous
> > > > >> > email,
> > > > >> > > but didn't get reply). Is there any plot in METviwer
that can
> > > > compare
> > > > >> the
> > > > >> > > wind direction b/t obs. and forecast?
> > > > >> > >
> > > > >> > >
> > > > >> > > Thank you.
> > > > >> > >
> > > > >> > > Binyu
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> > >
> > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway via
RT <
> > > > >> > > met_help at ucar.edu>
> > > > >> > > wrote:
> > > > >> > >
> > > > >> > > > Hi Binyu,
> > > > >> > > >
> > > > >> > > > The wind_thresh is a filtering threshold which is
applied to
> > the
> > > > >> wind
> > > > >> > > > SPEED. While the individual U and V components can be
> positive
> > > or
> > > > >> > > negative,
> > > > >> > > > the wind speed is always >= 0.
> > > > >> > > >
> > > > >> > > > The wind_thresh applies to the processing of the
VL1L2,
> > VAL1L2,
> > > > and
> > > > >> > VCNT
> > > > >> > > > line types which process U/V pairs. Let's say you've
set:
> > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > >> > > >
> > > > >> > > > Instead of seeing only 1 VL1L2 output line, you'll
get 4...
> > one
> > > > for
> > > > >> > each
> > > > >> > > > wind_thresh chosen.
> > > > >> > > > The first one (wind_thresh = NA) will include all U/V
pairs
> in
> > > the
> > > > >> > > masking
> > > > >> > > > area.
> > > > >> > > > The second one (wind_thresh ge1) will only include
those U/V
> > > pairs
> > > > >> > whose
> > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > >> > > > The third and fourth ones (wind_thresh ge5 and ge10)
will
> only
> > > > >> include
> > > > >> > > > those U/V pairs whose corresponding wind speed is at
least 5
> > m/s
> > > > or
> > > > >> > > 10m/s,
> > > > >> > > > respectively.
> > > > >> > > >
> > > > >> > > > The idea here is checking to see how the model
performs
> light
> > > and
> > > > >> > > variable
> > > > >> > > > winds (i.e. low wind speed) versus well-defined winds
(i.e.
> > high
> > > > >> wind
> > > > >> > > > speed).
> > > > >> > > >
> > > > >> > > > Thanks,
> > > > >> > > > John
> > > > >> > > >
> > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov
via RT
> <
> > > > >> > > > met_help at ucar.edu> wrote:
> > > > >> > > >
> > > > >> > > > >
> > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted
upon.
> > > > >> > > > > Transaction: Ticket created by binyu.wang at noaa.gov
> > > > >> > > > >        Queue: met_help
> > > > >> > > > >      Subject: wind_thresh
> > > > >> > > > >        Owner: Nobody
> > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > >> > > > >       Status: new
> > > > >> > > > >  Ticket <URL:
> > > > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > >> > > >
> > > > >> > > > >
> > > > >> > > > >
> > > > >> > > > > Hello,
> > > > >> > > > >
> > > > >> > > > > I have a question about setting up wind_thresh in
config
> > file:
> > > > >> > because
> > > > >> > > > the
> > > > >> > > > > wind can be negative and positive, so how to set
them
> > > correctly?
> > > > >> > > > > e.g, which is correct:
> > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
>=5.0&&<=10
> > > > >> ||>=-10&&<=-5,
> > > > >> > > > > >=10.0||<=-10 ];
> > > > >> > > > > or
> > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20  ];
> > > > >> > > > >
> > > > >> > > > > Thank you.
> > > > >> > > > > Binyu
> > > > >> > > > >
> > > > >> > > > >
> > > > >> > > >
> > > > >> > > >
> > > > >> > >
> > > > >> > >
> > > > >> >
> > > > >> >
> > > > >>
> > > > >>
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Thu Aug 27 15:49:13 2020

John,
Thank you. More questions here:
1. I finished running the verification using 9.1_beta2. Should I rerun
them
using 9.1? Will there be any difference for the point_stat output
between using the different versions?

2. How did you do the reorder?  Here is the script I used to do
asc2nc, I
used the default config file.
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/asc2grid_uwyo_SKTQ.sh

On Thu, Aug 27, 2020 at 5:00 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Binyu,
>
> Yes, I was just writing up this response...
>
> OK, I logged onto WCOSS and was able to replicate the behavior you
> describe.
>
> Here's what I found...
>
> met-9.1_beta2 does run with this data but met-9.1 does not. But
that's
> because there is a bug in met-9.1_beta2 that was fixed in met-9.1.
>
> Take a look at this line in point_stat.cc:
>
>
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861
>
> The bug in beta2 caused this logic not be enforced because "is_vgrd"
never
> evaluated to true.
>
> But the real point here is that by convention, MET expects that when
> processing winds as vectors, each point observation of UGRD is
immediately
> followed by a VGRD observation at the same level. We enforce this
> requirement to avoid the need of wasting processing time searching
through
> all the U/V pairs to match them up. In the data you passed to
point_stat
> your have 175 observations of UGRD followed by 175 observations of
VGRD. It
> just so happens that the order of the levels in both is identical...
but
> met-9.1_beta2 didn't actually check that to be sure! So if you made
a
> mistake in that regard, met-9.1_beta2 would not have caught it.
>
> To test things out, I reordered your data to interleave UGRD and
VGRD for
> each level. And the resulting file ran fine through point_stat. I've
> attached that NetCDF file here.
>
> Hope that helps clarify.
>
> Thanks,
> John
>
>
> On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> > Hello John,
> >
> > Just want to remind that I am still waiting for the answer. I am
sorry to
> > keep asking questions: I guess I am the one with most of the
questions.
> >
> >
> >
> > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Binyu,
> > >
> > > I believe that WCOSS is down this week and unavailable... at
least on
> the
> > > development side.
> > >
> > > If you're able to grab that sample data and post it on the RAL
> anonymous
> > > ftp site, I could grab it from there for testing.
> > >
> > > If that possible, please follow the instructions posted here:
> > >
> > >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> > >
> > > Thanks,
> > > John Halley Gotway
> > >
> > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> > > >
> > > > Hello again,
> > > >
> > > > Following your suggestion, I tried to look for more sites for
> > > verification
> > > > by using prepbufr data. Unfortunately I always got "0 matching
pair"
> > > when I
> > > > applied PB2NC and ensemble_stat. Here is the directory:
> > > >
> > > > cd
> > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > > >
> > > > % ./test_bufr.ecf >& bufr.log3
> > > >
> > > >
> > > >  I tried to diagnose the problem by converting the prepbufr
data to
> > grid
> > > > data using point_to_grid, the script is:
> > > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > > >
> > > > One of the output is:
> > > >
> > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > > regrid.prebufr.nc.txt
> > > >
> > > > It seems the PRES field does not look correct to me: it is
over
> > ~20000mb,
> > > > which is impossible. I confirmed with prepbufr people, the
unit
> inside
> > of
> > > > the prepbufr is mb, and the magnitude should be <1000mb. But I
can
> not
> > > find
> > > > where the error is. Any suggestions?
> > > >
> > > > Thank you.
> > > > Binyu
> > > >
> > > >
> > > >
> > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > > To keep these issues straight, I just created a separate RT
ticket
> > for
> > > > > Tatiana to look at. Hopefully you received a copy of it.
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <
> johnhg at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > > Binyu,
> > > > > >
> > > > > > I'm guessing that the output is not actually the same. In
the
> VL1L2
> > > and
> > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns should
now
> > > indicate
> > > > > that
> > > > > > you're using different thresholds: '>=1.0' should now be
changed
> to
> > > > > > '<=-1.0||>=1.0'
> > > > > >
> > > > > > You should confirm that those have changed to be confident
that
> you
> > > > > > actually ran the modified configuration. Once you've
confirmed
> > that,
> > > if
> > > > > the
> > > > > > counts and stats are still the same, that's really just a
> function
> > of
> > > > > your
> > > > > > data... about which I have no idea.
> > > > > >
> > > > > > John
> > > > > >
> > > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via RT
<
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > >>
> > > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > > >>
> > > > > >> John,
> > > > > >>
> > > > > >> Should I send another ticket to "Tatiana" about my log
error to
> > > upload
> > > > > >> point_stat*stat files to METViwer?
> > > > > >>
> > > > > >> I just tried the new config file (cat_thresh = [ <=-
1.0||>=1.0,
> > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as using
the old
> > > config
> > > > > >> file (
> > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have too
few
> > > matching
> > > > > >> pairs
> > > > > >> (only one or two pairs)?
> > > > > >>
> > > > > >> Binyu
> > > > > >>
> > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via RT
<
> > > > > >> met_help at ucar.edu>
> > > > > >> wrote:
> > > > > >>
> > > > > >> > Binyu,
> > > > > >> >
> > > > > >> > You're right, the cat_thresh for UGRD of >=1.0 would
only
> > > threshold
> > > > > the
> > > > > >> > positive values. If you want to apply both positive and
> negative
> > > you
> > > > > >> could
> > > > > >> > use:
> > > > > >> >
> > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > > >> >
> > > > > >> > Table 7.19 in the MET User's Guide describes the
contents of
> the
> > > > VCNT
> > > > > >> line
> > > > > >> > type:
> > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > > >> >
> > > > > >> > Looking at the description of those statistics, it
looks like
> > you
> > > > may
> > > > > be
> > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
DIR_ABSERR.
> > And
> > > > > >> > METviewer should be able to plot those statistics from
the
> VCNT
> > > line
> > > > > >> type.
> > > > > >> > That being said, I have never actually done this myself
for a
> > > model
> > > > > >> > evaluation.
> > > > > >> >
> > > > > >> > If METviewer isn't loading VCNT lines or plotting the
> statistics
> > > > from
> > > > > >> them,
> > > > > >> > we'll need to let Tatiana Burek know so that she can
fix it.
> > > > > >> >
> > > > > >> > Thanks,
> > > > > >> > John
> > > > > >> >
> > > > > >> >
> > > > > >> >
> > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov via
RT <
> > > > > >> > met_help at ucar.edu> wrote:
> > > > > >> >
> > > > > >> > >
> > > > > >> > > <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > >
> > > > > >> > >
> > > > > >> > > John,
> > > > > >> > >
> > > > > >> > > That makes sense for wind SPEED. Please look at the
> point_stat
> > > > > config:
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> >
> > > > > >>
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > > >> > > PointStatConfig
> > > > > >> > >
> > > > > >> > > field = [
> > > > > >> > >
> > > > > >> > >      {
> > > > > >> > >
> > > > > >> > >         name       = "UGRD";
> > > > > >> > >
> > > > > >> > >         level      = [ "P250" ];
> > > > > >> > >
> > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > > >> > >
> > > > > >> > >       },
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > the cat_thresh ONLY has positive too, does that mean
both
> > > > directions
> > > > > >> will
> > > > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > One more question here: following your suggestion, I
did
> > create
> > > > > VL1L2
> > > > > >> and
> > > > > >> > > VCNT,but I got strange error (
> > > > > >> > >
> > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > > >> when I
> > > > > >> > > tried to load to METviewer ( I mentioned that in one
of my
> > > > previous
> > > > > >> > email,
> > > > > >> > > but didn't get reply). Is there any plot in METviwer
that
> can
> > > > > compare
> > > > > >> the
> > > > > >> > > wind direction b/t obs. and forecast?
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > Thank you.
> > > > > >> > >
> > > > > >> > > Binyu
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway
via RT <
> > > > > >> > > met_help at ucar.edu>
> > > > > >> > > wrote:
> > > > > >> > >
> > > > > >> > > > Hi Binyu,
> > > > > >> > > >
> > > > > >> > > > The wind_thresh is a filtering threshold which is
applied
> to
> > > the
> > > > > >> wind
> > > > > >> > > > SPEED. While the individual U and V components can
be
> > positive
> > > > or
> > > > > >> > > negative,
> > > > > >> > > > the wind speed is always >= 0.
> > > > > >> > > >
> > > > > >> > > > The wind_thresh applies to the processing of the
VL1L2,
> > > VAL1L2,
> > > > > and
> > > > > >> > VCNT
> > > > > >> > > > line types which process U/V pairs. Let's say
you've set:
> > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > > >> > > >
> > > > > >> > > > Instead of seeing only 1 VL1L2 output line, you'll
get
> 4...
> > > one
> > > > > for
> > > > > >> > each
> > > > > >> > > > wind_thresh chosen.
> > > > > >> > > > The first one (wind_thresh = NA) will include all
U/V
> pairs
> > in
> > > > the
> > > > > >> > > masking
> > > > > >> > > > area.
> > > > > >> > > > The second one (wind_thresh ge1) will only include
those
> U/V
> > > > pairs
> > > > > >> > whose
> > > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > > >> > > > The third and fourth ones (wind_thresh ge5 and
ge10) will
> > only
> > > > > >> include
> > > > > >> > > > those U/V pairs whose corresponding wind speed is
at
> least 5
> > > m/s
> > > > > or
> > > > > >> > > 10m/s,
> > > > > >> > > > respectively.
> > > > > >> > > >
> > > > > >> > > > The idea here is checking to see how the model
performs
> > light
> > > > and
> > > > > >> > > variable
> > > > > >> > > > winds (i.e. low wind speed) versus well-defined
winds
> (i.e.
> > > high
> > > > > >> wind
> > > > > >> > > > speed).
> > > > > >> > > >
> > > > > >> > > > Thanks,
> > > > > >> > > > John
> > > > > >> > > >
> > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM binyu.wang at noaa.gov
via
> RT
> > <
> > > > > >> > > > met_help at ucar.edu> wrote:
> > > > > >> > > >
> > > > > >> > > > >
> > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was acted
upon.
> > > > > >> > > > > Transaction: Ticket created by
binyu.wang at noaa.gov
> > > > > >> > > > >        Queue: met_help
> > > > > >> > > > >      Subject: wind_thresh
> > > > > >> > > > >        Owner: Nobody
> > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > > >> > > > >       Status: new
> > > > > >> > > > >  Ticket <URL:
> > > > > >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > > >> > > >
> > > > > >> > > > >
> > > > > >> > > > >
> > > > > >> > > > > Hello,
> > > > > >> > > > >
> > > > > >> > > > > I have a question about setting up wind_thresh in
config
> > > file:
> > > > > >> > because
> > > > > >> > > > the
> > > > > >> > > > > wind can be negative and positive, so how to set
them
> > > > correctly?
> > > > > >> > > > > e.g, which is correct:
> > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
>=5.0&&<=10
> > > > > >> ||>=-10&&<=-5,
> > > > > >> > > > > >=10.0||<=-10 ];
> > > > > >> > > > > or
> > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20
];
> > > > > >> > > > >
> > > > > >> > > > > Thank you.
> > > > > >> > > > > Binyu
> > > > > >> > > > >
> > > > > >> > > > >
> > > > > >> > > >
> > > > > >> > > >
> > > > > >> > >
> > > > > >> > >
> > > > > >> >
> > > > > >> >
> > > > > >>
> > > > > >>
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Thu Aug 27 16:34:49 2020

Should you rerun with met-9.1? Generally, the beta releases are meant
for
testing new development, not production runs. I don't know of any
results
off the top of my head that would be different between beta2 and 9.1,
but
there certainly could be something. If you'd like to use beta2, it'd
be
prudent to carefully review the release notes and consider which
changes
may or may not impact your results. Or you could rerun everything with
met-9.1. Or you could just rerun a couple of example cases, diff your
results with beta2 output, and decide what you'd like to do from
there.

Listed below are the steps I ran to reorder your data. But I
definitely
don't recommend this as a general purpose way of doing this. Instead,
it's
just a hack I used for testing, and it was only for U/V obs, not other
obs
in the file.

# Run Rscript to dump existing NetCDF file to ascii (probably won't
work on
wcoss because the ncdf4 library probably doesn't exist)

Rscript
/Volumes/d1/projects/MET/MET_development/MET-
develop/met/scripts/Rscripts/pntnc2ascii.R
asc2nc_PANO_20200203t12Z.nc > asc2nc_PANO_20200203t12Z.txt

# Subset the UGRD and VGRD obs into separate files

grep UGRD asc2nc_PANO_20200203t12Z.txt >
asc2nc_PANO_20200203t12Z_UGRD.txt

grep VGRD asc2nc_PANO_20200203t12Z.txt >
asc2nc_PANO_20200203t12Z_VGRD.txt

# Run the paste command to stitch them back together

paste -d
"\n" asc2nc_PANO_20200203t12Z_UGRD.txt
asc2nc_PANO_20200203t12Z_VGRD.txt
> asc2nc_PANO_20200203t12Z_sorted.txt

# Run ascii2nc to regenerate NetCDF output

ascii2nc asc2nc_PANO_20200203t12Z_sorted.txt
asc2nc_PANO_20200203t12Z_sorted.nc

On Thu, Aug 27, 2020 at 3:49 PM binyu.wang at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> John,
> Thank you. More questions here:
> 1. I finished running the verification using 9.1_beta2. Should I
rerun them
> using 9.1? Will there be any difference for the point_stat output
> between using the different versions?
>
> 2. How did you do the reorder?  Here is the script I used to do
asc2nc, I
> used the default config file.
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/asc2grid_uwyo_SKTQ.sh
>
> On Thu, Aug 27, 2020 at 5:00 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Binyu,
> >
> > Yes, I was just writing up this response...
> >
> > OK, I logged onto WCOSS and was able to replicate the behavior you
> > describe.
> >
> > Here's what I found...
> >
> > met-9.1_beta2 does run with this data but met-9.1 does not. But
that's
> > because there is a bug in met-9.1_beta2 that was fixed in met-9.1.
> >
> > Take a look at this line in point_stat.cc:
> >
> >
>
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861
> >
> > The bug in beta2 caused this logic not be enforced because
"is_vgrd"
> never
> > evaluated to true.
> >
> > But the real point here is that by convention, MET expects that
when
> > processing winds as vectors, each point observation of UGRD is
> immediately
> > followed by a VGRD observation at the same level. We enforce this
> > requirement to avoid the need of wasting processing time searching
> through
> > all the U/V pairs to match them up. In the data you passed to
point_stat
> > your have 175 observations of UGRD followed by 175 observations of
VGRD.
> It
> > just so happens that the order of the levels in both is
identical... but
> > met-9.1_beta2 didn't actually check that to be sure! So if you
made a
> > mistake in that regard, met-9.1_beta2 would not have caught it.
> >
> > To test things out, I reordered your data to interleave UGRD and
VGRD for
> > each level. And the resulting file ran fine through point_stat.
I've
> > attached that NetCDF file here.
> >
> > Hope that helps clarify.
> >
> > Thanks,
> > John
> >
> >
> > On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > >
> > > Hello John,
> > >
> > > Just want to remind that I am still waiting for the answer. I am
sorry
> to
> > > keep asking questions: I guess I am the one with most of the
questions.
> > >
> > >
> > >
> > > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > > Binyu,
> > > >
> > > > I believe that WCOSS is down this week and unavailable... at
least on
> > the
> > > > development side.
> > > >
> > > > If you're able to grab that sample data and post it on the RAL
> > anonymous
> > > > ftp site, I could grab it from there for testing.
> > > >
> > > > If that possible, please follow the instructions posted here:
> > > >
> > > >
> > >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> > > >
> > > > Thanks,
> > > > John Halley Gotway
> > > >
> > > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> > > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > >
> > > > > Hello again,
> > > > >
> > > > > Following your suggestion, I tried to look for more sites
for
> > > > verification
> > > > > by using prepbufr data. Unfortunately I always got "0
matching
> pair"
> > > > when I
> > > > > applied PB2NC and ensemble_stat. Here is the directory:
> > > > >
> > > > > cd
> > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > > > >
> > > > > % ./test_bufr.ecf >& bufr.log3
> > > > >
> > > > >
> > > > >  I tried to diagnose the problem by converting the prepbufr
data to
> > > grid
> > > > > data using point_to_grid, the script is:
> > > > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > > > >
> > > > > One of the output is:
> > > > >
> > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > > > regrid.prebufr.nc.txt
> > > > >
> > > > > It seems the PRES field does not look correct to me: it is
over
> > > ~20000mb,
> > > > > which is impossible. I confirmed with prepbufr people, the
unit
> > inside
> > > of
> > > > > the prepbufr is mb, and the magnitude should be <1000mb. But
I can
> > not
> > > > find
> > > > > where the error is. Any suggestions?
> > > > >
> > > > > Thank you.
> > > > > Binyu
> > > > >
> > > > >
> > > > >
> > > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT <
> > > > > met_help at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > > To keep these issues straight, I just created a separate
RT
> ticket
> > > for
> > > > > > Tatiana to look at. Hopefully you received a copy of it.
> > > > > >
> > > > > > Thanks,
> > > > > > John
> > > > > >
> > > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <
> > johnhg at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > > Binyu,
> > > > > > >
> > > > > > > I'm guessing that the output is not actually the same.
In the
> > VL1L2
> > > > and
> > > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns
should now
> > > > indicate
> > > > > > that
> > > > > > > you're using different thresholds: '>=1.0' should now be
> changed
> > to
> > > > > > > '<=-1.0||>=1.0'
> > > > > > >
> > > > > > > You should confirm that those have changed to be
confident that
> > you
> > > > > > > actually ran the modified configuration. Once you've
confirmed
> > > that,
> > > > if
> > > > > > the
> > > > > > > counts and stats are still the same, that's really just
a
> > function
> > > of
> > > > > > your
> > > > > > > data... about which I have no idea.
> > > > > > >
> > > > > > > John
> > > > > > >
> > > > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov via
RT <
> > > > > > > met_help at ucar.edu> wrote:
> > > > > > >
> > > > > > >>
> > > > > > >> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >
> > > > > > >>
> > > > > > >> John,
> > > > > > >>
> > > > > > >> Should I send another ticket to "Tatiana" about my log
error
> to
> > > > upload
> > > > > > >> point_stat*stat files to METViwer?
> > > > > > >>
> > > > > > >> I just tried the new config file (cat_thresh = [
> <=-1.0||>=1.0,
> > > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as using
the old
> > > > config
> > > > > > >> file (
> > > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have
too few
> > > > matching
> > > > > > >> pairs
> > > > > > >> (only one or two pairs)?
> > > > > > >>
> > > > > > >> Binyu
> > > > > > >>
> > > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway via
RT <
> > > > > > >> met_help at ucar.edu>
> > > > > > >> wrote:
> > > > > > >>
> > > > > > >> > Binyu,
> > > > > > >> >
> > > > > > >> > You're right, the cat_thresh for UGRD of >=1.0 would
only
> > > > threshold
> > > > > > the
> > > > > > >> > positive values. If you want to apply both positive
and
> > negative
> > > > you
> > > > > > >> could
> > > > > > >> > use:
> > > > > > >> >
> > > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > > > >> >
> > > > > > >> > Table 7.19 in the MET User's Guide describes the
contents of
> > the
> > > > > VCNT
> > > > > > >> line
> > > > > > >> > type:
> > > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > > > >> >
> > > > > > >> > Looking at the description of those statistics, it
looks
> like
> > > you
> > > > > may
> > > > > > be
> > > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
> DIR_ABSERR.
> > > And
> > > > > > >> > METviewer should be able to plot those statistics
from the
> > VCNT
> > > > line
> > > > > > >> type.
> > > > > > >> > That being said, I have never actually done this
myself for
> a
> > > > model
> > > > > > >> > evaluation.
> > > > > > >> >
> > > > > > >> > If METviewer isn't loading VCNT lines or plotting the
> > statistics
> > > > > from
> > > > > > >> them,
> > > > > > >> > we'll need to let Tatiana Burek know so that she can
fix it.
> > > > > > >> >
> > > > > > >> > Thanks,
> > > > > > >> > John
> > > > > > >> >
> > > > > > >> >
> > > > > > >> >
> > > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov
via RT
> <
> > > > > > >> > met_help at ucar.edu> wrote:
> > > > > > >> >
> > > > > > >> > >
> > > > > > >> > > <URL:
> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > >
> > > > > > >> > >
> > > > > > >> > > John,
> > > > > > >> > >
> > > > > > >> > > That makes sense for wind SPEED. Please look at the
> > point_stat
> > > > > > config:
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> >
> > > > > > >>
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > > > >> > > PointStatConfig
> > > > > > >> > >
> > > > > > >> > > field = [
> > > > > > >> > >
> > > > > > >> > >      {
> > > > > > >> > >
> > > > > > >> > >         name       = "UGRD";
> > > > > > >> > >
> > > > > > >> > >         level      = [ "P250" ];
> > > > > > >> > >
> > > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > > > >> > >
> > > > > > >> > >       },
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > > the cat_thresh ONLY has positive too, does that
mean both
> > > > > directions
> > > > > > >> will
> > > > > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > > One more question here: following your suggestion,
I did
> > > create
> > > > > > VL1L2
> > > > > > >> and
> > > > > > >> > > VCNT,but I got strange error (
> > > > > > >> > >
> > > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > > > >> when I
> > > > > > >> > > tried to load to METviewer ( I mentioned that in
one of my
> > > > > previous
> > > > > > >> > email,
> > > > > > >> > > but didn't get reply). Is there any plot in
METviwer that
> > can
> > > > > > compare
> > > > > > >> the
> > > > > > >> > > wind direction b/t obs. and forecast?
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > > Thank you.
> > > > > > >> > >
> > > > > > >> > > Binyu
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley Gotway
via RT
> <
> > > > > > >> > > met_help at ucar.edu>
> > > > > > >> > > wrote:
> > > > > > >> > >
> > > > > > >> > > > Hi Binyu,
> > > > > > >> > > >
> > > > > > >> > > > The wind_thresh is a filtering threshold which is
> applied
> > to
> > > > the
> > > > > > >> wind
> > > > > > >> > > > SPEED. While the individual U and V components
can be
> > > positive
> > > > > or
> > > > > > >> > > negative,
> > > > > > >> > > > the wind speed is always >= 0.
> > > > > > >> > > >
> > > > > > >> > > > The wind_thresh applies to the processing of the
VL1L2,
> > > > VAL1L2,
> > > > > > and
> > > > > > >> > VCNT
> > > > > > >> > > > line types which process U/V pairs. Let's say
you've
> set:
> > > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > > > >> > > >
> > > > > > >> > > > Instead of seeing only 1 VL1L2 output line,
you'll get
> > 4...
> > > > one
> > > > > > for
> > > > > > >> > each
> > > > > > >> > > > wind_thresh chosen.
> > > > > > >> > > > The first one (wind_thresh = NA) will include all
U/V
> > pairs
> > > in
> > > > > the
> > > > > > >> > > masking
> > > > > > >> > > > area.
> > > > > > >> > > > The second one (wind_thresh ge1) will only
include those
> > U/V
> > > > > pairs
> > > > > > >> > whose
> > > > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > > > >> > > > The third and fourth ones (wind_thresh ge5 and
ge10)
> will
> > > only
> > > > > > >> include
> > > > > > >> > > > those U/V pairs whose corresponding wind speed is
at
> > least 5
> > > > m/s
> > > > > > or
> > > > > > >> > > 10m/s,
> > > > > > >> > > > respectively.
> > > > > > >> > > >
> > > > > > >> > > > The idea here is checking to see how the model
performs
> > > light
> > > > > and
> > > > > > >> > > variable
> > > > > > >> > > > winds (i.e. low wind speed) versus well-defined
winds
> > (i.e.
> > > > high
> > > > > > >> wind
> > > > > > >> > > > speed).
> > > > > > >> > > >
> > > > > > >> > > > Thanks,
> > > > > > >> > > > John
> > > > > > >> > > >
> > > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM
binyu.wang at noaa.gov via
> > RT
> > > <
> > > > > > >> > > > met_help at ucar.edu> wrote:
> > > > > > >> > > >
> > > > > > >> > > > >
> > > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was
acted
> upon.
> > > > > > >> > > > > Transaction: Ticket created by
binyu.wang at noaa.gov
> > > > > > >> > > > >        Queue: met_help
> > > > > > >> > > > >      Subject: wind_thresh
> > > > > > >> > > > >        Owner: Nobody
> > > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > > > >> > > > >       Status: new
> > > > > > >> > > > >  Ticket <URL:
> > > > > > >> >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > > > >> > > >
> > > > > > >> > > > >
> > > > > > >> > > > >
> > > > > > >> > > > > Hello,
> > > > > > >> > > > >
> > > > > > >> > > > > I have a question about setting up wind_thresh
in
> config
> > > > file:
> > > > > > >> > because
> > > > > > >> > > > the
> > > > > > >> > > > > wind can be negative and positive, so how to
set them
> > > > > correctly?
> > > > > > >> > > > > e.g, which is correct:
> > > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
>=5.0&&<=10
> > > > > > >> ||>=-10&&<=-5,
> > > > > > >> > > > > >=10.0||<=-10 ];
> > > > > > >> > > > > or
> > > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,  >=20
];
> > > > > > >> > > > >
> > > > > > >> > > > > Thank you.
> > > > > > >> > > > > Binyu
> > > > > > >> > > > >
> > > > > > >> > > > >
> > > > > > >> > > >
> > > > > > >> > > >
> > > > > > >> > >
> > > > > > >> > >
> > > > > > >> >
> > > > > > >> >
> > > > > > >>
> > > > > > >>
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Thu Aug 27 17:00:11 2020

Got it, thank you very much!
Before you are very tired of answering my questions, there is one more
from
two weeks ago:

 I tried to do gefs verification by using prepbufr data. Unfortunately
I
always got "0 matching pair" when I applied PB2NC and ensemble_stat.
Here
is the directory:

cd
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf

% ./test_bufr.ecf >& bufr.log3

 I tried to diagnose the problem by converting the prepbufr data to
grid
data using point_to_grid, the script is:
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh

One of the output is:

/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
regrid.prebufr.nc.txt

It seems the PRES field does not look correct to me: it is over
~20000mb,
which is impossible. I confirmed with prepbufr people, the unit inside
of
the prepbufr is mb, and the magnitude should be <1000mb. But I can not
find
where the error is. Any suggestions?

John Halley Gotway via RT <met_help at ucar.edu>

On Thu, Aug 27, 2020 at 6:34 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Should you rerun with met-9.1? Generally, the beta releases are
meant for
> testing new development, not production runs. I don't know of any
results
> off the top of my head that would be different between beta2 and
9.1, but
> there certainly could be something. If you'd like to use beta2, it'd
be
> prudent to carefully review the release notes and consider which
changes
> may or may not impact your results. Or you could rerun everything
with
> met-9.1. Or you could just rerun a couple of example cases, diff
your
> results with beta2 output, and decide what you'd like to do from
there.
>
> Listed below are the steps I ran to reorder your data. But I
definitely
> don't recommend this as a general purpose way of doing this.
Instead, it's
> just a hack I used for testing, and it was only for U/V obs, not
other obs
> in the file.
>
> # Run Rscript to dump existing NetCDF file to ascii (probably won't
work on
> wcoss because the ncdf4 library probably doesn't exist)
>
> Rscript
>
> /Volumes/d1/projects/MET/MET_development/MET-
develop/met/scripts/Rscripts/pntnc2ascii.R
> asc2nc_PANO_20200203t12Z.nc > asc2nc_PANO_20200203t12Z.txt
>
>
> # Subset the UGRD and VGRD obs into separate files
>
> grep UGRD asc2nc_PANO_20200203t12Z.txt >
asc2nc_PANO_20200203t12Z_UGRD.txt
>
> grep VGRD asc2nc_PANO_20200203t12Z.txt >
asc2nc_PANO_20200203t12Z_VGRD.txt
>
>
> # Run the paste command to stitch them back together
>
> paste -d
> "\n" asc2nc_PANO_20200203t12Z_UGRD.txt
asc2nc_PANO_20200203t12Z_VGRD.txt
> > asc2nc_PANO_20200203t12Z_sorted.txt
>
>
> # Run ascii2nc to regenerate NetCDF output
>
> ascii2nc asc2nc_PANO_20200203t12Z_sorted.txt
asc2nc_PANO_20200203t12Z_
> sorted.nc
>
> On Thu, Aug 27, 2020 at 3:49 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> > John,
> > Thank you. More questions here:
> > 1. I finished running the verification using 9.1_beta2. Should I
rerun
> them
> > using 9.1? Will there be any difference for the point_stat output
> > between using the different versions?
> >
> > 2. How did you do the reorder?  Here is the script I used to do
asc2nc, I
> > used the default config file.
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/asc2grid_uwyo_SKTQ.sh
> >
> > On Thu, Aug 27, 2020 at 5:00 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Binyu,
> > >
> > > Yes, I was just writing up this response...
> > >
> > > OK, I logged onto WCOSS and was able to replicate the behavior
you
> > > describe.
> > >
> > > Here's what I found...
> > >
> > > met-9.1_beta2 does run with this data but met-9.1 does not. But
that's
> > > because there is a bug in met-9.1_beta2 that was fixed in met-
9.1.
> > >
> > > Take a look at this line in point_stat.cc:
> > >
> > >
> >
>
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861
> > >
> > > The bug in beta2 caused this logic not be enforced because
"is_vgrd"
> > never
> > > evaluated to true.
> > >
> > > But the real point here is that by convention, MET expects that
when
> > > processing winds as vectors, each point observation of UGRD is
> > immediately
> > > followed by a VGRD observation at the same level. We enforce
this
> > > requirement to avoid the need of wasting processing time
searching
> > through
> > > all the U/V pairs to match them up. In the data you passed to
> point_stat
> > > your have 175 observations of UGRD followed by 175 observations
of
> VGRD.
> > It
> > > just so happens that the order of the levels in both is
identical...
> but
> > > met-9.1_beta2 didn't actually check that to be sure! So if you
made a
> > > mistake in that regard, met-9.1_beta2 would not have caught it.
> > >
> > > To test things out, I reordered your data to interleave UGRD and
VGRD
> for
> > > each level. And the resulting file ran fine through point_stat.
I've
> > > attached that NetCDF file here.
> > >
> > > Hope that helps clarify.
> > >
> > > Thanks,
> > > John
> > >
> > >
> > > On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> > > >
> > > > Hello John,
> > > >
> > > > Just want to remind that I am still waiting for the answer. I
am
> sorry
> > to
> > > > keep asking questions: I guess I am the one with most of the
> questions.
> > > >
> > > >
> > > >
> > > > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > > Binyu,
> > > > >
> > > > > I believe that WCOSS is down this week and unavailable... at
least
> on
> > > the
> > > > > development side.
> > > > >
> > > > > If you're able to grab that sample data and post it on the
RAL
> > > anonymous
> > > > > ftp site, I could grab it from there for testing.
> > > > >
> > > > > If that possible, please follow the instructions posted
here:
> > > > >
> > > > >
> > > >
> > >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> > > > >
> > > > > Thanks,
> > > > > John Halley Gotway
> > > > >
> > > > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> > > > > met_help at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > > >
> > > > > > Hello again,
> > > > > >
> > > > > > Following your suggestion, I tried to look for more sites
for
> > > > > verification
> > > > > > by using prepbufr data. Unfortunately I always got "0
matching
> > pair"
> > > > > when I
> > > > > > applied PB2NC and ensemble_stat. Here is the directory:
> > > > > >
> > > > > > cd
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > > > > >
> > > > > > % ./test_bufr.ecf >& bufr.log3
> > > > > >
> > > > > >
> > > > > >  I tried to diagnose the problem by converting the
prepbufr data
> to
> > > > grid
> > > > > > data using point_to_grid, the script is:
> > > > > >
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > > > > >
> > > > > > One of the output is:
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > > > > regrid.prebufr.nc.txt
> > > > > >
> > > > > > It seems the PRES field does not look correct to me: it is
over
> > > > ~20000mb,
> > > > > > which is impossible. I confirmed with prepbufr people, the
unit
> > > inside
> > > > of
> > > > > > the prepbufr is mb, and the magnitude should be <1000mb.
But I
> can
> > > not
> > > > > find
> > > > > > where the error is. Any suggestions?
> > > > > >
> > > > > > Thank you.
> > > > > > Binyu
> > > > > >
> > > > > >
> > > > > >
> > > > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > > To keep these issues straight, I just created a separate
RT
> > ticket
> > > > for
> > > > > > > Tatiana to look at. Hopefully you received a copy of it.
> > > > > > >
> > > > > > > Thanks,
> > > > > > > John
> > > > > > >
> > > > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <
> > > johnhg at ucar.edu>
> > > > > > > wrote:
> > > > > > >
> > > > > > > > Binyu,
> > > > > > > >
> > > > > > > > I'm guessing that the output is not actually the same.
In the
> > > VL1L2
> > > > > and
> > > > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns
should now
> > > > > indicate
> > > > > > > that
> > > > > > > > you're using different thresholds: '>=1.0' should now
be
> > changed
> > > to
> > > > > > > > '<=-1.0||>=1.0'
> > > > > > > >
> > > > > > > > You should confirm that those have changed to be
confident
> that
> > > you
> > > > > > > > actually ran the modified configuration. Once you've
> confirmed
> > > > that,
> > > > > if
> > > > > > > the
> > > > > > > > counts and stats are still the same, that's really
just a
> > > function
> > > > of
> > > > > > > your
> > > > > > > > data... about which I have no idea.
> > > > > > > >
> > > > > > > > John
> > > > > > > >
> > > > > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov
via RT <
> > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > >
> > > > > > > >>
> > > > > > > >> <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > >
> > > > > > > >>
> > > > > > > >> John,
> > > > > > > >>
> > > > > > > >> Should I send another ticket to "Tatiana" about my
log error
> > to
> > > > > upload
> > > > > > > >> point_stat*stat files to METViwer?
> > > > > > > >>
> > > > > > > >> I just tried the new config file (cat_thresh = [
> > <=-1.0||>=1.0,
> > > > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as
using the
> old
> > > > > config
> > > > > > > >> file (
> > > > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have
too few
> > > > > matching
> > > > > > > >> pairs
> > > > > > > >> (only one or two pairs)?
> > > > > > > >>
> > > > > > > >> Binyu
> > > > > > > >>
> > > > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway
via RT <
> > > > > > > >> met_help at ucar.edu>
> > > > > > > >> wrote:
> > > > > > > >>
> > > > > > > >> > Binyu,
> > > > > > > >> >
> > > > > > > >> > You're right, the cat_thresh for UGRD of >=1.0
would only
> > > > > threshold
> > > > > > > the
> > > > > > > >> > positive values. If you want to apply both positive
and
> > > negative
> > > > > you
> > > > > > > >> could
> > > > > > > >> > use:
> > > > > > > >> >
> > > > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > > > > >> >
> > > > > > > >> > Table 7.19 in the MET User's Guide describes the
contents
> of
> > > the
> > > > > > VCNT
> > > > > > > >> line
> > > > > > > >> > type:
> > > > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > > > > >> >
> > > > > > > >> > Looking at the description of those statistics, it
looks
> > like
> > > > you
> > > > > > may
> > > > > > > be
> > > > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
> > DIR_ABSERR.
> > > > And
> > > > > > > >> > METviewer should be able to plot those statistics
from the
> > > VCNT
> > > > > line
> > > > > > > >> type.
> > > > > > > >> > That being said, I have never actually done this
myself
> for
> > a
> > > > > model
> > > > > > > >> > evaluation.
> > > > > > > >> >
> > > > > > > >> > If METviewer isn't loading VCNT lines or plotting
the
> > > statistics
> > > > > > from
> > > > > > > >> them,
> > > > > > > >> > we'll need to let Tatiana Burek know so that she
can fix
> it.
> > > > > > > >> >
> > > > > > > >> > Thanks,
> > > > > > > >> > John
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov
via
> RT
> > <
> > > > > > > >> > met_help at ucar.edu> wrote:
> > > > > > > >> >
> > > > > > > >> > >
> > > > > > > >> > > <URL:
> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > >
> > > > > > > >> > >
> > > > > > > >> > > John,
> > > > > > > >> > >
> > > > > > > >> > > That makes sense for wind SPEED. Please look at
the
> > > point_stat
> > > > > > > config:
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> >
> > > > > > > >>
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > > > > >> > > PointStatConfig
> > > > > > > >> > >
> > > > > > > >> > > field = [
> > > > > > > >> > >
> > > > > > > >> > >      {
> > > > > > > >> > >
> > > > > > > >> > >         name       = "UGRD";
> > > > > > > >> > >
> > > > > > > >> > >         level      = [ "P250" ];
> > > > > > > >> > >
> > > > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > > > > >> > >
> > > > > > > >> > >       },
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > the cat_thresh ONLY has positive too, does that
mean
> both
> > > > > > directions
> > > > > > > >> will
> > > > > > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > One more question here: following your
suggestion, I did
> > > > create
> > > > > > > VL1L2
> > > > > > > >> and
> > > > > > > >> > > VCNT,but I got strange error (
> > > > > > > >> > >
> > > > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > > > > >> when I
> > > > > > > >> > > tried to load to METviewer ( I mentioned that in
one of
> my
> > > > > > previous
> > > > > > > >> > email,
> > > > > > > >> > > but didn't get reply). Is there any plot in
METviwer
> that
> > > can
> > > > > > > compare
> > > > > > > >> the
> > > > > > > >> > > wind direction b/t obs. and forecast?
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > Thank you.
> > > > > > > >> > >
> > > > > > > >> > > Binyu
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley
Gotway via
> RT
> > <
> > > > > > > >> > > met_help at ucar.edu>
> > > > > > > >> > > wrote:
> > > > > > > >> > >
> > > > > > > >> > > > Hi Binyu,
> > > > > > > >> > > >
> > > > > > > >> > > > The wind_thresh is a filtering threshold which
is
> > applied
> > > to
> > > > > the
> > > > > > > >> wind
> > > > > > > >> > > > SPEED. While the individual U and V components
can be
> > > > positive
> > > > > > or
> > > > > > > >> > > negative,
> > > > > > > >> > > > the wind speed is always >= 0.
> > > > > > > >> > > >
> > > > > > > >> > > > The wind_thresh applies to the processing of
the
> VL1L2,
> > > > > VAL1L2,
> > > > > > > and
> > > > > > > >> > VCNT
> > > > > > > >> > > > line types which process U/V pairs. Let's say
you've
> > set:
> > > > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > > > > >> > > >
> > > > > > > >> > > > Instead of seeing only 1 VL1L2 output line,
you'll get
> > > 4...
> > > > > one
> > > > > > > for
> > > > > > > >> > each
> > > > > > > >> > > > wind_thresh chosen.
> > > > > > > >> > > > The first one (wind_thresh = NA) will include
all U/V
> > > pairs
> > > > in
> > > > > > the
> > > > > > > >> > > masking
> > > > > > > >> > > > area.
> > > > > > > >> > > > The second one (wind_thresh ge1) will only
include
> those
> > > U/V
> > > > > > pairs
> > > > > > > >> > whose
> > > > > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > > > > >> > > > The third and fourth ones (wind_thresh ge5 and
ge10)
> > will
> > > > only
> > > > > > > >> include
> > > > > > > >> > > > those U/V pairs whose corresponding wind speed
is at
> > > least 5
> > > > > m/s
> > > > > > > or
> > > > > > > >> > > 10m/s,
> > > > > > > >> > > > respectively.
> > > > > > > >> > > >
> > > > > > > >> > > > The idea here is checking to see how the model
> performs
> > > > light
> > > > > > and
> > > > > > > >> > > variable
> > > > > > > >> > > > winds (i.e. low wind speed) versus well-defined
winds
> > > (i.e.
> > > > > high
> > > > > > > >> wind
> > > > > > > >> > > > speed).
> > > > > > > >> > > >
> > > > > > > >> > > > Thanks,
> > > > > > > >> > > > John
> > > > > > > >> > > >
> > > > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM
binyu.wang at noaa.gov
> via
> > > RT
> > > > <
> > > > > > > >> > > > met_help at ucar.edu> wrote:
> > > > > > > >> > > >
> > > > > > > >> > > > >
> > > > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was
acted
> > upon.
> > > > > > > >> > > > > Transaction: Ticket created by
binyu.wang at noaa.gov
> > > > > > > >> > > > >        Queue: met_help
> > > > > > > >> > > > >      Subject: wind_thresh
> > > > > > > >> > > > >        Owner: Nobody
> > > > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > > > > >> > > > >       Status: new
> > > > > > > >> > > > >  Ticket <URL:
> > > > > > > >> >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > > > > >> > > >
> > > > > > > >> > > > >
> > > > > > > >> > > > >
> > > > > > > >> > > > > Hello,
> > > > > > > >> > > > >
> > > > > > > >> > > > > I have a question about setting up
wind_thresh in
> > config
> > > > > file:
> > > > > > > >> > because
> > > > > > > >> > > > the
> > > > > > > >> > > > > wind can be negative and positive, so how to
set
> them
> > > > > > correctly?
> > > > > > > >> > > > > e.g, which is correct:
> > > > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
> >=5.0&&<=10
> > > > > > > >> ||>=-10&&<=-5,
> > > > > > > >> > > > > >=10.0||<=-10 ];
> > > > > > > >> > > > > or
> > > > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,
>=20  ];
> > > > > > > >> > > > >
> > > > > > > >> > > > > Thank you.
> > > > > > > >> > > > > Binyu
> > > > > > > >> > > > >
> > > > > > > >> > > > >
> > > > > > > >> > > >
> > > > > > > >> > > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >>
> > > > > > > >>
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Thu Aug 27 17:26:52 2020

Sorry, I'm headed out for the day. But will keep this message marked
unread
and get back to it when I get a chance.

Thanks,
John

On Thu, Aug 27, 2020 at 5:00 PM binyu.wang at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> Got it, thank you very much!
> Before you are very tired of answering my questions, there is one
more from
> two weeks ago:
>
>  I tried to do gefs verification by using prepbufr data.
Unfortunately I
> always got "0 matching pair" when I applied PB2NC and ensemble_stat.
Here
> is the directory:
>
> cd
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
>
> % ./test_bufr.ecf >& bufr.log3
>
>
>  I tried to diagnose the problem by converting the prepbufr data to
grid
> data using point_to_grid, the script is:
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
>
> One of the output is:
>
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> regrid.prebufr.nc.txt
>
> It seems the PRES field does not look correct to me: it is over
~20000mb,
> which is impossible. I confirmed with prepbufr people, the unit
inside of
> the prepbufr is mb, and the magnitude should be <1000mb. But I can
not find
> where the error is. Any suggestions?
>
>
> John Halley Gotway via RT <met_help at ucar.edu>
>
> On Thu, Aug 27, 2020 at 6:34 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Should you rerun with met-9.1? Generally, the beta releases are
meant for
> > testing new development, not production runs. I don't know of any
results
> > off the top of my head that would be different between beta2 and
9.1, but
> > there certainly could be something. If you'd like to use beta2,
it'd be
> > prudent to carefully review the release notes and consider which
changes
> > may or may not impact your results. Or you could rerun everything
with
> > met-9.1. Or you could just rerun a couple of example cases, diff
your
> > results with beta2 output, and decide what you'd like to do from
there.
> >
> > Listed below are the steps I ran to reorder your data. But I
definitely
> > don't recommend this as a general purpose way of doing this.
Instead,
> it's
> > just a hack I used for testing, and it was only for U/V obs, not
other
> obs
> > in the file.
> >
> > # Run Rscript to dump existing NetCDF file to ascii (probably
won't work
> on
> > wcoss because the ncdf4 library probably doesn't exist)
> >
> > Rscript
> >
> >
> /Volumes/d1/projects/MET/MET_development/MET-
develop/met/scripts/Rscripts/pntnc2ascii.R
> > asc2nc_PANO_20200203t12Z.nc > asc2nc_PANO_20200203t12Z.txt
> >
> >
> > # Subset the UGRD and VGRD obs into separate files
> >
> > grep UGRD asc2nc_PANO_20200203t12Z.txt >
> asc2nc_PANO_20200203t12Z_UGRD.txt
> >
> > grep VGRD asc2nc_PANO_20200203t12Z.txt >
> asc2nc_PANO_20200203t12Z_VGRD.txt
> >
> >
> > # Run the paste command to stitch them back together
> >
> > paste -d
> > "\n" asc2nc_PANO_20200203t12Z_UGRD.txt
asc2nc_PANO_20200203t12Z_VGRD.txt
> > > asc2nc_PANO_20200203t12Z_sorted.txt
> >
> >
> > # Run ascii2nc to regenerate NetCDF output
> >
> > ascii2nc asc2nc_PANO_20200203t12Z_sorted.txt
asc2nc_PANO_20200203t12Z_
> > sorted.nc
> >
> > On Thu, Aug 27, 2020 at 3:49 PM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > >
> > > John,
> > > Thank you. More questions here:
> > > 1. I finished running the verification using 9.1_beta2. Should I
rerun
> > them
> > > using 9.1? Will there be any difference for the point_stat
output
> > > between using the different versions?
> > >
> > > 2. How did you do the reorder?  Here is the script I used to do
> asc2nc, I
> > > used the default config file.
> > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/asc2grid_uwyo_SKTQ.sh
> > >
> > > On Thu, Aug 27, 2020 at 5:00 PM John Halley Gotway via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > > Binyu,
> > > >
> > > > Yes, I was just writing up this response...
> > > >
> > > > OK, I logged onto WCOSS and was able to replicate the behavior
you
> > > > describe.
> > > >
> > > > Here's what I found...
> > > >
> > > > met-9.1_beta2 does run with this data but met-9.1 does not.
But
> that's
> > > > because there is a bug in met-9.1_beta2 that was fixed in met-
9.1.
> > > >
> > > > Take a look at this line in point_stat.cc:
> > > >
> > > >
> > >
> >
>
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861
> > > >
> > > > The bug in beta2 caused this logic not be enforced because
"is_vgrd"
> > > never
> > > > evaluated to true.
> > > >
> > > > But the real point here is that by convention, MET expects
that when
> > > > processing winds as vectors, each point observation of UGRD is
> > > immediately
> > > > followed by a VGRD observation at the same level. We enforce
this
> > > > requirement to avoid the need of wasting processing time
searching
> > > through
> > > > all the U/V pairs to match them up. In the data you passed to
> > point_stat
> > > > your have 175 observations of UGRD followed by 175
observations of
> > VGRD.
> > > It
> > > > just so happens that the order of the levels in both is
identical...
> > but
> > > > met-9.1_beta2 didn't actually check that to be sure! So if you
made a
> > > > mistake in that regard, met-9.1_beta2 would not have caught
it.
> > > >
> > > > To test things out, I reordered your data to interleave UGRD
and VGRD
> > for
> > > > each level. And the resulting file ran fine through
point_stat. I've
> > > > attached that NetCDF file here.
> > > >
> > > > Hope that helps clarify.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > >
> > > > On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > >
> > > > > Hello John,
> > > > >
> > > > > Just want to remind that I am still waiting for the answer.
I am
> > sorry
> > > to
> > > > > keep asking questions: I guess I am the one with most of the
> > questions.
> > > > >
> > > > >
> > > > >
> > > > > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> > > > > met_help at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > > Binyu,
> > > > > >
> > > > > > I believe that WCOSS is down this week and unavailable...
at
> least
> > on
> > > > the
> > > > > > development side.
> > > > > >
> > > > > > If you're able to grab that sample data and post it on the
RAL
> > > > anonymous
> > > > > > ftp site, I could grab it from there for testing.
> > > > > >
> > > > > > If that possible, please follow the instructions posted
here:
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> > > > > >
> > > > > > Thanks,
> > > > > > John Halley Gotway
> > > > > >
> > > > > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT
<
> > > > > > met_help at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > >
> > > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >
> > > > > > >
> > > > > > > Hello again,
> > > > > > >
> > > > > > > Following your suggestion, I tried to look for more
sites for
> > > > > > verification
> > > > > > > by using prepbufr data. Unfortunately I always got "0
matching
> > > pair"
> > > > > > when I
> > > > > > > applied PB2NC and ensemble_stat. Here is the directory:
> > > > > > >
> > > > > > > cd
> > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > > > > > >
> > > > > > > % ./test_bufr.ecf >& bufr.log3
> > > > > > >
> > > > > > >
> > > > > > >  I tried to diagnose the problem by converting the
prepbufr
> data
> > to
> > > > > grid
> > > > > > > data using point_to_grid, the script is:
> > > > > > >
> > /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > > > > > >
> > > > > > > One of the output is:
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > > > > > regrid.prebufr.nc.txt
> > > > > > >
> > > > > > > It seems the PRES field does not look correct to me: it
is over
> > > > > ~20000mb,
> > > > > > > which is impossible. I confirmed with prepbufr people,
the unit
> > > > inside
> > > > > of
> > > > > > > the prepbufr is mb, and the magnitude should be <1000mb.
But I
> > can
> > > > not
> > > > > > find
> > > > > > > where the error is. Any suggestions?
> > > > > > >
> > > > > > > Thank you.
> > > > > > > Binyu
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via
RT <
> > > > > > > met_help at ucar.edu>
> > > > > > > wrote:
> > > > > > >
> > > > > > > > To keep these issues straight, I just created a
separate RT
> > > ticket
> > > > > for
> > > > > > > > Tatiana to look at. Hopefully you received a copy of
it.
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > > John
> > > > > > > >
> > > > > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <
> > > > johnhg at ucar.edu>
> > > > > > > > wrote:
> > > > > > > >
> > > > > > > > > Binyu,
> > > > > > > > >
> > > > > > > > > I'm guessing that the output is not actually the
same. In
> the
> > > > VL1L2
> > > > > > and
> > > > > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns
should
> now
> > > > > > indicate
> > > > > > > > that
> > > > > > > > > you're using different thresholds: '>=1.0' should
now be
> > > changed
> > > > to
> > > > > > > > > '<=-1.0||>=1.0'
> > > > > > > > >
> > > > > > > > > You should confirm that those have changed to be
confident
> > that
> > > > you
> > > > > > > > > actually ran the modified configuration. Once you've
> > confirmed
> > > > > that,
> > > > > > if
> > > > > > > > the
> > > > > > > > > counts and stats are still the same, that's really
just a
> > > > function
> > > > > of
> > > > > > > > your
> > > > > > > > > data... about which I have no idea.
> > > > > > > > >
> > > > > > > > > John
> > > > > > > > >
> > > > > > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov
via
> RT <
> > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > >
> > > > > > > > >>
> > > > > > > > >> <URL:
> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > >
> > > > > > > > >>
> > > > > > > > >> John,
> > > > > > > > >>
> > > > > > > > >> Should I send another ticket to "Tatiana" about my
log
> error
> > > to
> > > > > > upload
> > > > > > > > >> point_stat*stat files to METViwer?
> > > > > > > > >>
> > > > > > > > >> I just tried the new config file (cat_thresh = [
> > > <=-1.0||>=1.0,
> > > > > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as
using the
> > old
> > > > > > config
> > > > > > > > >> file (
> > > > > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I
have too
> few
> > > > > > matching
> > > > > > > > >> pairs
> > > > > > > > >> (only one or two pairs)?
> > > > > > > > >>
> > > > > > > > >> Binyu
> > > > > > > > >>
> > > > > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway
via RT
> <
> > > > > > > > >> met_help at ucar.edu>
> > > > > > > > >> wrote:
> > > > > > > > >>
> > > > > > > > >> > Binyu,
> > > > > > > > >> >
> > > > > > > > >> > You're right, the cat_thresh for UGRD of >=1.0
would
> only
> > > > > > threshold
> > > > > > > > the
> > > > > > > > >> > positive values. If you want to apply both
positive and
> > > > negative
> > > > > > you
> > > > > > > > >> could
> > > > > > > > >> > use:
> > > > > > > > >> >
> > > > > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > > > > > >> >
> > > > > > > > >> > Table 7.19 in the MET User's Guide describes the
> contents
> > of
> > > > the
> > > > > > > VCNT
> > > > > > > > >> line
> > > > > > > > >> > type:
> > > > > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > > > > > >> >
> > > > > > > > >> > Looking at the description of those statistics,
it looks
> > > like
> > > > > you
> > > > > > > may
> > > > > > > > be
> > > > > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR,
and
> > > DIR_ABSERR.
> > > > > And
> > > > > > > > >> > METviewer should be able to plot those statistics
from
> the
> > > > VCNT
> > > > > > line
> > > > > > > > >> type.
> > > > > > > > >> > That being said, I have never actually done this
myself
> > for
> > > a
> > > > > > model
> > > > > > > > >> > evaluation.
> > > > > > > > >> >
> > > > > > > > >> > If METviewer isn't loading VCNT lines or plotting
the
> > > > statistics
> > > > > > > from
> > > > > > > > >> them,
> > > > > > > > >> > we'll need to let Tatiana Burek know so that she
can fix
> > it.
> > > > > > > > >> >
> > > > > > > > >> > Thanks,
> > > > > > > > >> > John
> > > > > > > > >> >
> > > > > > > > >> >
> > > > > > > > >> >
> > > > > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM
binyu.wang at noaa.gov via
> > RT
> > > <
> > > > > > > > >> > met_help at ucar.edu> wrote:
> > > > > > > > >> >
> > > > > > > > >> > >
> > > > > > > > >> > > <URL:
> > > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > > >
> > > > > > > > >> > >
> > > > > > > > >> > > John,
> > > > > > > > >> > >
> > > > > > > > >> > > That makes sense for wind SPEED. Please look at
the
> > > > point_stat
> > > > > > > > config:
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> >
> > > > > > > > >>
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > > > > > >> > > PointStatConfig
> > > > > > > > >> > >
> > > > > > > > >> > > field = [
> > > > > > > > >> > >
> > > > > > > > >> > >      {
> > > > > > > > >> > >
> > > > > > > > >> > >         name       = "UGRD";
> > > > > > > > >> > >
> > > > > > > > >> > >         level      = [ "P250" ];
> > > > > > > > >> > >
> > > > > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > > > > > >> > >
> > > > > > > > >> > >       },
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> > > the cat_thresh ONLY has positive too, does that
mean
> > both
> > > > > > > directions
> > > > > > > > >> will
> > > > > > > > >> > > be considered? (>= refers either ge 5 or le
-5)?
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> > > One more question here: following your
suggestion, I
> did
> > > > > create
> > > > > > > > VL1L2
> > > > > > > > >> and
> > > > > > > > >> > > VCNT,but I got strange error (
> > > > > > > > >> > >
> > > > > >
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > > > > > >> when I
> > > > > > > > >> > > tried to load to METviewer ( I mentioned that
in one
> of
> > my
> > > > > > > previous
> > > > > > > > >> > email,
> > > > > > > > >> > > but didn't get reply). Is there any plot in
METviwer
> > that
> > > > can
> > > > > > > > compare
> > > > > > > > >> the
> > > > > > > > >> > > wind direction b/t obs. and forecast?
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> > > Thank you.
> > > > > > > > >> > >
> > > > > > > > >> > > Binyu
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley
Gotway via
> > RT
> > > <
> > > > > > > > >> > > met_help at ucar.edu>
> > > > > > > > >> > > wrote:
> > > > > > > > >> > >
> > > > > > > > >> > > > Hi Binyu,
> > > > > > > > >> > > >
> > > > > > > > >> > > > The wind_thresh is a filtering threshold
which is
> > > applied
> > > > to
> > > > > > the
> > > > > > > > >> wind
> > > > > > > > >> > > > SPEED. While the individual U and V
components can
> be
> > > > > positive
> > > > > > > or
> > > > > > > > >> > > negative,
> > > > > > > > >> > > > the wind speed is always >= 0.
> > > > > > > > >> > > >
> > > > > > > > >> > > > The wind_thresh applies to the processing of
the
> > VL1L2,
> > > > > > VAL1L2,
> > > > > > > > and
> > > > > > > > >> > VCNT
> > > > > > > > >> > > > line types which process U/V pairs. Let's say
you've
> > > set:
> > > > > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > > > > > >> > > >
> > > > > > > > >> > > > Instead of seeing only 1 VL1L2 output line,
you'll
> get
> > > > 4...
> > > > > > one
> > > > > > > > for
> > > > > > > > >> > each
> > > > > > > > >> > > > wind_thresh chosen.
> > > > > > > > >> > > > The first one (wind_thresh = NA) will include
all
> U/V
> > > > pairs
> > > > > in
> > > > > > > the
> > > > > > > > >> > > masking
> > > > > > > > >> > > > area.
> > > > > > > > >> > > > The second one (wind_thresh ge1) will only
include
> > those
> > > > U/V
> > > > > > > pairs
> > > > > > > > >> > whose
> > > > > > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > > > > > >> > > > The third and fourth ones (wind_thresh ge5
and ge10)
> > > will
> > > > > only
> > > > > > > > >> include
> > > > > > > > >> > > > those U/V pairs whose corresponding wind
speed is at
> > > > least 5
> > > > > > m/s
> > > > > > > > or
> > > > > > > > >> > > 10m/s,
> > > > > > > > >> > > > respectively.
> > > > > > > > >> > > >
> > > > > > > > >> > > > The idea here is checking to see how the
model
> > performs
> > > > > light
> > > > > > > and
> > > > > > > > >> > > variable
> > > > > > > > >> > > > winds (i.e. low wind speed) versus well-
defined
> winds
> > > > (i.e.
> > > > > > high
> > > > > > > > >> wind
> > > > > > > > >> > > > speed).
> > > > > > > > >> > > >
> > > > > > > > >> > > > Thanks,
> > > > > > > > >> > > > John
> > > > > > > > >> > > >
> > > > > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM
binyu.wang at noaa.gov
> > via
> > > > RT
> > > > > <
> > > > > > > > >> > > > met_help at ucar.edu> wrote:
> > > > > > > > >> > > >
> > > > > > > > >> > > > >
> > > > > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was
acted
> > > upon.
> > > > > > > > >> > > > > Transaction: Ticket created by
> binyu.wang at noaa.gov
> > > > > > > > >> > > > >        Queue: met_help
> > > > > > > > >> > > > >      Subject: wind_thresh
> > > > > > > > >> > > > >        Owner: Nobody
> > > > > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > > > > > >> > > > >       Status: new
> > > > > > > > >> > > > >  Ticket <URL:
> > > > > > > > >> >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > > > > > >> > > >
> > > > > > > > >> > > > >
> > > > > > > > >> > > > >
> > > > > > > > >> > > > > Hello,
> > > > > > > > >> > > > >
> > > > > > > > >> > > > > I have a question about setting up
wind_thresh in
> > > config
> > > > > > file:
> > > > > > > > >> > because
> > > > > > > > >> > > > the
> > > > > > > > >> > > > > wind can be negative and positive, so how
to set
> > them
> > > > > > > correctly?
> > > > > > > > >> > > > > e.g, which is correct:
> > > > > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
> > >=5.0&&<=10
> > > > > > > > >> ||>=-10&&<=-5,
> > > > > > > > >> > > > > >=10.0||<=-10 ];
> > > > > > > > >> > > > > or
> > > > > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,
>=20
> ];
> > > > > > > > >> > > > >
> > > > > > > > >> > > > > Thank you.
> > > > > > > > >> > > > > Binyu
> > > > > > > > >> > > > >
> > > > > > > > >> > > > >
> > > > > > > > >> > > >
> > > > > > > > >> > > >
> > > > > > > > >> > >
> > > > > > > > >> > >
> > > > > > > > >> >
> > > > > > > > >> >
> > > > > > > > >>
> > > > > > > > >>
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Fri Aug 28 14:28:47 2020

Binyu,

OK I'm taking a look in this log file:
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf/bufr.log3

Whenever folks get 0 matched pairs from Point-Stat, I always advise
that
they increase the verbosity level to list out counts of reason codes
explaining why obs were or were not used, such as below:

DEBUG 3: Number of matched pairs   = 132

DEBUG 3: Observations processed    = 89893

DEBUG 3: Rejected: station id      = 0

DEBUG 3: Rejected: obs type        = 79788

DEBUG 3: Rejected: valid time      = 0

DEBUG 3: Rejected: bad obs value   = 0

DEBUG 3: Rejected: off the grid    = 5

DEBUG 3: Rejected: topography      = 0

DEBUG 3: Rejected: level mismatch  = 9219

DEBUG 3: Rejected: quality marker  = 0

DEBUG 3: Rejected: message type    = 344

DEBUG 3: Rejected: masking region  = 405

DEBUG 3: Rejected: bad fcst value  = 0

DEBUG 3: Rejected: bad climo mean  = 0

DEBUG 3: Rejected: bad climo stdev = 0

DEBUG 3: Rejected: duplicates      = 0

Unfortunately though, Ensemble-Stat does NOT track and report the same
counts!

So I wanted to actually run Point-Stat using 1 of the ensemble members
with
those same point observations at verbosity level 3.

BUT when I logged on WCOSS, I could not actually locate those files.
So you
could either...

(1) Do this testing yourself. Run Point-Stat to compare 1 ensemble
member
to these obs and check the verbosity level 3 output to see why there
are 0
matched pairs.

(2) Or you could post the following files to our anonymous FTP site
following these instructions:

http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp

I'd need:

(1) Ens file:

gefs01.t00z.grid3.f12.grib2

(2) Obs file:

/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/
gdas.2020020300.prebufr.f00.nc

(3) Ensemble-Stat config file:

/gpfs/dell1/ptmp/Binyu.Wang/gens/tmpnwprd/verf_g2o_gefs_00_dev.25142/verf_g2o_ens_gens_config.bufr

Please just let me know.

Thanks,
John

On Thu, Aug 27, 2020 at 5:26 PM John Halley Gotway <johnhg at ucar.edu>
wrote:

> Sorry, I'm headed out for the day. But will keep this message marked
> unread and get back to it when I get a chance.
>
> Thanks,
> John
>
> On Thu, Aug 27, 2020 at 5:00 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>>
>> Got it, thank you very much!
>> Before you are very tired of answering my questions, there is one
more
>> from
>> two weeks ago:
>>
>>  I tried to do gefs verification by using prepbufr data.
Unfortunately I
>> always got "0 matching pair" when I applied PB2NC and
ensemble_stat. Here
>> is the directory:
>>
>> cd
>>
>>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
>>
>> % ./test_bufr.ecf >& bufr.log3
>>
>>
>>  I tried to diagnose the problem by converting the prepbufr data to
grid
>> data using point_to_grid, the script is:
>> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
>>
>> One of the output is:
>>
>>
>>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
>> regrid.prebufr.nc.txt
>>
>> It seems the PRES field does not look correct to me: it is over
~20000mb,
>> which is impossible. I confirmed with prepbufr people, the unit
inside of
>> the prepbufr is mb, and the magnitude should be <1000mb. But I can
not
>> find
>> where the error is. Any suggestions?
>>
>>
>> John Halley Gotway via RT <met_help at ucar.edu>
>>
>> On Thu, Aug 27, 2020 at 6:34 PM John Halley Gotway via RT <
>> met_help at ucar.edu>
>> wrote:
>>
>> > Should you rerun with met-9.1? Generally, the beta releases are
meant
>> for
>> > testing new development, not production runs. I don't know of any
>> results
>> > off the top of my head that would be different between beta2 and
9.1,
>> but
>> > there certainly could be something. If you'd like to use beta2,
it'd be
>> > prudent to carefully review the release notes and consider which
changes
>> > may or may not impact your results. Or you could rerun everything
with
>> > met-9.1. Or you could just rerun a couple of example cases, diff
your
>> > results with beta2 output, and decide what you'd like to do from
there.
>> >
>> > Listed below are the steps I ran to reorder your data. But I
definitely
>> > don't recommend this as a general purpose way of doing this.
Instead,
>> it's
>> > just a hack I used for testing, and it was only for U/V obs, not
other
>> obs
>> > in the file.
>> >
>> > # Run Rscript to dump existing NetCDF file to ascii (probably
won't
>> work on
>> > wcoss because the ncdf4 library probably doesn't exist)
>> >
>> > Rscript
>> >
>> >
>> /Volumes/d1/projects/MET/MET_development/MET-
develop/met/scripts/Rscripts/pntnc2ascii.R
>> > asc2nc_PANO_20200203t12Z.nc > asc2nc_PANO_20200203t12Z.txt
>> >
>> >
>> > # Subset the UGRD and VGRD obs into separate files
>> >
>> > grep UGRD asc2nc_PANO_20200203t12Z.txt >
>> asc2nc_PANO_20200203t12Z_UGRD.txt
>> >
>> > grep VGRD asc2nc_PANO_20200203t12Z.txt >
>> asc2nc_PANO_20200203t12Z_VGRD.txt
>> >
>> >
>> > # Run the paste command to stitch them back together
>> >
>> > paste -d
>> > "\n" asc2nc_PANO_20200203t12Z_UGRD.txt
asc2nc_PANO_20200203t12Z_VGRD.txt
>> > > asc2nc_PANO_20200203t12Z_sorted.txt
>> >
>> >
>> > # Run ascii2nc to regenerate NetCDF output
>> >
>> > ascii2nc asc2nc_PANO_20200203t12Z_sorted.txt
asc2nc_PANO_20200203t12Z_
>> > sorted.nc
>> >
>> > On Thu, Aug 27, 2020 at 3:49 PM binyu.wang at noaa.gov via RT <
>> > met_help at ucar.edu> wrote:
>> >
>> > >
>> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>> > >
>> > > John,
>> > > Thank you. More questions here:
>> > > 1. I finished running the verification using 9.1_beta2. Should
I rerun
>> > them
>> > > using 9.1? Will there be any difference for the point_stat
output
>> > > between using the different versions?
>> > >
>> > > 2. How did you do the reorder?  Here is the script I used to do
>> asc2nc, I
>> > > used the default config file.
>> > >
>> > >
>> >
>>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/asc2grid_uwyo_SKTQ.sh
>> > >
>> > > On Thu, Aug 27, 2020 at 5:00 PM John Halley Gotway via RT <
>> > > met_help at ucar.edu>
>> > > wrote:
>> > >
>> > > > Binyu,
>> > > >
>> > > > Yes, I was just writing up this response...
>> > > >
>> > > > OK, I logged onto WCOSS and was able to replicate the
behavior you
>> > > > describe.
>> > > >
>> > > > Here's what I found...
>> > > >
>> > > > met-9.1_beta2 does run with this data but met-9.1 does not.
But
>> that's
>> > > > because there is a bug in met-9.1_beta2 that was fixed in
met-9.1.
>> > > >
>> > > > Take a look at this line in point_stat.cc:
>> > > >
>> > > >
>> > >
>> >
>>
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861
>> > > >
>> > > > The bug in beta2 caused this logic not be enforced because
"is_vgrd"
>> > > never
>> > > > evaluated to true.
>> > > >
>> > > > But the real point here is that by convention, MET expects
that when
>> > > > processing winds as vectors, each point observation of UGRD
is
>> > > immediately
>> > > > followed by a VGRD observation at the same level. We enforce
this
>> > > > requirement to avoid the need of wasting processing time
searching
>> > > through
>> > > > all the U/V pairs to match them up. In the data you passed to
>> > point_stat
>> > > > your have 175 observations of UGRD followed by 175
observations of
>> > VGRD.
>> > > It
>> > > > just so happens that the order of the levels in both is
identical...
>> > but
>> > > > met-9.1_beta2 didn't actually check that to be sure! So if
you made
>> a
>> > > > mistake in that regard, met-9.1_beta2 would not have caught
it.
>> > > >
>> > > > To test things out, I reordered your data to interleave UGRD
and
>> VGRD
>> > for
>> > > > each level. And the resulting file ran fine through
point_stat. I've
>> > > > attached that NetCDF file here.
>> > > >
>> > > > Hope that helps clarify.
>> > > >
>> > > > Thanks,
>> > > > John
>> > > >
>> > > >
>> > > > On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT <
>> > > > met_help at ucar.edu> wrote:
>> > > >
>> > > > >
>> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>> > > > >
>> > > > > Hello John,
>> > > > >
>> > > > > Just want to remind that I am still waiting for the answer.
I am
>> > sorry
>> > > to
>> > > > > keep asking questions: I guess I am the one with most of
the
>> > questions.
>> > > > >
>> > > > >
>> > > > >
>> > > > > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
>> > > > > met_help at ucar.edu>
>> > > > > wrote:
>> > > > >
>> > > > > > Binyu,
>> > > > > >
>> > > > > > I believe that WCOSS is down this week and unavailable...
at
>> least
>> > on
>> > > > the
>> > > > > > development side.
>> > > > > >
>> > > > > > If you're able to grab that sample data and post it on
the RAL
>> > > > anonymous
>> > > > > > ftp site, I could grab it from there for testing.
>> > > > > >
>> > > > > > If that possible, please follow the instructions posted
here:
>> > > > > >
>> > > > > >
>> > > > >
>> > > >
>> > >
>> >
>> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
>> > > > > >
>> > > > > > Thanks,
>> > > > > > John Halley Gotway
>> > > > > >
>> > > > > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT
<
>> > > > > > met_help at ucar.edu>
>> > > > > > wrote:
>> > > > > >
>> > > > > > >
>> > > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>> >
>> > > > > > >
>> > > > > > > Hello again,
>> > > > > > >
>> > > > > > > Following your suggestion, I tried to look for more
sites for
>> > > > > > verification
>> > > > > > > by using prepbufr data. Unfortunately I always got "0
matching
>> > > pair"
>> > > > > > when I
>> > > > > > > applied PB2NC and ensemble_stat. Here is the directory:
>> > > > > > >
>> > > > > > > cd
>> > > > > > >
>> > > > > > >
>> > > > > >
>> > > > >
>> > > >
>> > >
>> >
>>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
>> > > > > > >
>> > > > > > > % ./test_bufr.ecf >& bufr.log3
>> > > > > > >
>> > > > > > >
>> > > > > > >  I tried to diagnose the problem by converting the
prepbufr
>> data
>> > to
>> > > > > grid
>> > > > > > > data using point_to_grid, the script is:
>> > > > > > >
>> >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
>> > > > > > >
>> > > > > > > One of the output is:
>> > > > > > >
>> > > > > > >
>> > > > > > >
>> > > > > >
>> > > > >
>> > > >
>> > >
>> >
>>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
>> > > > > > > regrid.prebufr.nc.txt
>> > > > > > >
>> > > > > > > It seems the PRES field does not look correct to me: it
is
>> over
>> > > > > ~20000mb,
>> > > > > > > which is impossible. I confirmed with prepbufr people,
the
>> unit
>> > > > inside
>> > > > > of
>> > > > > > > the prepbufr is mb, and the magnitude should be
<1000mb. But I
>> > can
>> > > > not
>> > > > > > find
>> > > > > > > where the error is. Any suggestions?
>> > > > > > >
>> > > > > > > Thank you.
>> > > > > > > Binyu
>> > > > > > >
>> > > > > > >
>> > > > > > >
>> > > > > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via
RT <
>> > > > > > > met_help at ucar.edu>
>> > > > > > > wrote:
>> > > > > > >
>> > > > > > > > To keep these issues straight, I just created a
separate RT
>> > > ticket
>> > > > > for
>> > > > > > > > Tatiana to look at. Hopefully you received a copy of
it.
>> > > > > > > >
>> > > > > > > > Thanks,
>> > > > > > > > John
>> > > > > > > >
>> > > > > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <
>> > > > johnhg at ucar.edu>
>> > > > > > > > wrote:
>> > > > > > > >
>> > > > > > > > > Binyu,
>> > > > > > > > >
>> > > > > > > > > I'm guessing that the output is not actually the
same. In
>> the
>> > > > VL1L2
>> > > > > > and
>> > > > > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns
should
>> now
>> > > > > > indicate
>> > > > > > > > that
>> > > > > > > > > you're using different thresholds: '>=1.0' should
now be
>> > > changed
>> > > > to
>> > > > > > > > > '<=-1.0||>=1.0'
>> > > > > > > > >
>> > > > > > > > > You should confirm that those have changed to be
confident
>> > that
>> > > > you
>> > > > > > > > > actually ran the modified configuration. Once
you've
>> > confirmed
>> > > > > that,
>> > > > > > if
>> > > > > > > > the
>> > > > > > > > > counts and stats are still the same, that's really
just a
>> > > > function
>> > > > > of
>> > > > > > > > your
>> > > > > > > > > data... about which I have no idea.
>> > > > > > > > >
>> > > > > > > > > John
>> > > > > > > > >
>> > > > > > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov
via
>> RT <
>> > > > > > > > > met_help at ucar.edu> wrote:
>> > > > > > > > >
>> > > > > > > > >>
>> > > > > > > > >> <URL:
>> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>> > > >
>> > > > > > > > >>
>> > > > > > > > >> John,
>> > > > > > > > >>
>> > > > > > > > >> Should I send another ticket to "Tatiana" about my
log
>> error
>> > > to
>> > > > > > upload
>> > > > > > > > >> point_stat*stat files to METViwer?
>> > > > > > > > >>
>> > > > > > > > >> I just tried the new config file (cat_thresh = [
>> > > <=-1.0||>=1.0,
>> > > > > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as
using
>> the
>> > old
>> > > > > > config
>> > > > > > > > >> file (
>> > > > > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I
have too
>> few
>> > > > > > matching
>> > > > > > > > >> pairs
>> > > > > > > > >> (only one or two pairs)?
>> > > > > > > > >>
>> > > > > > > > >> Binyu
>> > > > > > > > >>
>> > > > > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway
via
>> RT <
>> > > > > > > > >> met_help at ucar.edu>
>> > > > > > > > >> wrote:
>> > > > > > > > >>
>> > > > > > > > >> > Binyu,
>> > > > > > > > >> >
>> > > > > > > > >> > You're right, the cat_thresh for UGRD of >=1.0
would
>> only
>> > > > > > threshold
>> > > > > > > > the
>> > > > > > > > >> > positive values. If you want to apply both
positive and
>> > > > negative
>> > > > > > you
>> > > > > > > > >> could
>> > > > > > > > >> > use:
>> > > > > > > > >> >
>> > > > > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
>> > > > > > > > >> >
>> > > > > > > > >> > Table 7.19 in the MET User's Guide describes the
>> contents
>> > of
>> > > > the
>> > > > > > > VCNT
>> > > > > > > > >> line
>> > > > > > > > >> > type:
>> > > > > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47
PM.png]
>> > > > > > > > >> >
>> > > > > > > > >> > Looking at the description of those statistics,
it
>> looks
>> > > like
>> > > > > you
>> > > > > > > may
>> > > > > > > > be
>> > > > > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR,
and
>> > > DIR_ABSERR.
>> > > > > And
>> > > > > > > > >> > METviewer should be able to plot those
statistics from
>> the
>> > > > VCNT
>> > > > > > line
>> > > > > > > > >> type.
>> > > > > > > > >> > That being said, I have never actually done this
myself
>> > for
>> > > a
>> > > > > > model
>> > > > > > > > >> > evaluation.
>> > > > > > > > >> >
>> > > > > > > > >> > If METviewer isn't loading VCNT lines or
plotting the
>> > > > statistics
>> > > > > > > from
>> > > > > > > > >> them,
>> > > > > > > > >> > we'll need to let Tatiana Burek know so that she
can
>> fix
>> > it.
>> > > > > > > > >> >
>> > > > > > > > >> > Thanks,
>> > > > > > > > >> > John
>> > > > > > > > >> >
>> > > > > > > > >> >
>> > > > > > > > >> >
>> > > > > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM
binyu.wang at noaa.gov
>> via
>> > RT
>> > > <
>> > > > > > > > >> > met_help at ucar.edu> wrote:
>> > > > > > > > >> >
>> > > > > > > > >> > >
>> > > > > > > > >> > > <URL:
>> > > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>> > > > > >
>> > > > > > > > >> > >
>> > > > > > > > >> > > John,
>> > > > > > > > >> > >
>> > > > > > > > >> > > That makes sense for wind SPEED. Please look
at the
>> > > > point_stat
>> > > > > > > > config:
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> >
>> > > > > > > > >>
>> > > > > > > >
>> > > > > > >
>> > > > > >
>> > > > >
>> > > >
>> > >
>> >
>>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
>> > > > > > > > >> > > PointStatConfig
>> > > > > > > > >> > >
>> > > > > > > > >> > > field = [
>> > > > > > > > >> > >
>> > > > > > > > >> > >      {
>> > > > > > > > >> > >
>> > > > > > > > >> > >         name       = "UGRD";
>> > > > > > > > >> > >
>> > > > > > > > >> > >         level      = [ "P250" ];
>> > > > > > > > >> > >
>> > > > > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
>> > > > > > > > >> > >
>> > > > > > > > >> > >       },
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> > > the cat_thresh ONLY has positive too, does
that mean
>> > both
>> > > > > > > directions
>> > > > > > > > >> will
>> > > > > > > > >> > > be considered? (>= refers either ge 5 or le
-5)?
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> > > One more question here: following your
suggestion, I
>> did
>> > > > > create
>> > > > > > > > VL1L2
>> > > > > > > > >> and
>> > > > > > > > >> > > VCNT,but I got strange error (
>> > > > > > > > >> > >
>> > > > > >
>> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
>> > > > > > > > >> when I
>> > > > > > > > >> > > tried to load to METviewer ( I mentioned that
in one
>> of
>> > my
>> > > > > > > previous
>> > > > > > > > >> > email,
>> > > > > > > > >> > > but didn't get reply). Is there any plot in
METviwer
>> > that
>> > > > can
>> > > > > > > > compare
>> > > > > > > > >> the
>> > > > > > > > >> > > wind direction b/t obs. and forecast?
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> > > Thank you.
>> > > > > > > > >> > >
>> > > > > > > > >> > > Binyu
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley
Gotway
>> via
>> > RT
>> > > <
>> > > > > > > > >> > > met_help at ucar.edu>
>> > > > > > > > >> > > wrote:
>> > > > > > > > >> > >
>> > > > > > > > >> > > > Hi Binyu,
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > The wind_thresh is a filtering threshold
which is
>> > > applied
>> > > > to
>> > > > > > the
>> > > > > > > > >> wind
>> > > > > > > > >> > > > SPEED. While the individual U and V
components can
>> be
>> > > > > positive
>> > > > > > > or
>> > > > > > > > >> > > negative,
>> > > > > > > > >> > > > the wind speed is always >= 0.
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > The wind_thresh applies to the processing of
the
>> > VL1L2,
>> > > > > > VAL1L2,
>> > > > > > > > and
>> > > > > > > > >> > VCNT
>> > > > > > > > >> > > > line types which process U/V pairs. Let's
say
>> you've
>> > > set:
>> > > > > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > Instead of seeing only 1 VL1L2 output line,
you'll
>> get
>> > > > 4...
>> > > > > > one
>> > > > > > > > for
>> > > > > > > > >> > each
>> > > > > > > > >> > > > wind_thresh chosen.
>> > > > > > > > >> > > > The first one (wind_thresh = NA) will
include all
>> U/V
>> > > > pairs
>> > > > > in
>> > > > > > > the
>> > > > > > > > >> > > masking
>> > > > > > > > >> > > > area.
>> > > > > > > > >> > > > The second one (wind_thresh ge1) will only
include
>> > those
>> > > > U/V
>> > > > > > > pairs
>> > > > > > > > >> > whose
>> > > > > > > > >> > > > corresponding wind speed is at least 1 m/s.
>> > > > > > > > >> > > > The third and fourth ones (wind_thresh ge5
and
>> ge10)
>> > > will
>> > > > > only
>> > > > > > > > >> include
>> > > > > > > > >> > > > those U/V pairs whose corresponding wind
speed is
>> at
>> > > > least 5
>> > > > > > m/s
>> > > > > > > > or
>> > > > > > > > >> > > 10m/s,
>> > > > > > > > >> > > > respectively.
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > The idea here is checking to see how the
model
>> > performs
>> > > > > light
>> > > > > > > and
>> > > > > > > > >> > > variable
>> > > > > > > > >> > > > winds (i.e. low wind speed) versus well-
defined
>> winds
>> > > > (i.e.
>> > > > > > high
>> > > > > > > > >> wind
>> > > > > > > > >> > > > speed).
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > Thanks,
>> > > > > > > > >> > > > John
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM
>> binyu.wang at noaa.gov
>> > via
>> > > > RT
>> > > > > <
>> > > > > > > > >> > > > met_help at ucar.edu> wrote:
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > >
>> > > > > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074
was acted
>> > > upon.
>> > > > > > > > >> > > > > Transaction: Ticket created by
>> binyu.wang at noaa.gov
>> > > > > > > > >> > > > >        Queue: met_help
>> > > > > > > > >> > > > >      Subject: wind_thresh
>> > > > > > > > >> > > > >        Owner: Nobody
>> > > > > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
>> > > > > > > > >> > > > >       Status: new
>> > > > > > > > >> > > > >  Ticket <URL:
>> > > > > > > > >> >
>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>> > > > > > > > >> > > >
>> > > > > > > > >> > > > >
>> > > > > > > > >> > > > >
>> > > > > > > > >> > > > > Hello,
>> > > > > > > > >> > > > >
>> > > > > > > > >> > > > > I have a question about setting up
wind_thresh in
>> > > config
>> > > > > > file:
>> > > > > > > > >> > because
>> > > > > > > > >> > > > the
>> > > > > > > > >> > > > > wind can be negative and positive, so how
to set
>> > them
>> > > > > > > correctly?
>> > > > > > > > >> > > > > e.g, which is correct:
>> > > > > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
>> > >=5.0&&<=10
>> > > > > > > > >> ||>=-10&&<=-5,
>> > > > > > > > >> > > > > >=10.0||<=-10 ];
>> > > > > > > > >> > > > > or
>> > > > > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,
>=20
>> ];
>> > > > > > > > >> > > > >
>> > > > > > > > >> > > > > Thank you.
>> > > > > > > > >> > > > > Binyu
>> > > > > > > > >> > > > >
>> > > > > > > > >> > > > >
>> > > > > > > > >> > > >
>> > > > > > > > >> > > >
>> > > > > > > > >> > >
>> > > > > > > > >> > >
>> > > > > > > > >> >
>> > > > > > > > >> >
>> > > > > > > > >>
>> > > > > > > > >>
>> > > > > > > >
>> > > > > > > >
>> > > > > > >
>> > > > > > >
>> > > > > >
>> > > > > >
>> > > > >
>> > > > >
>> > > >
>> > > >
>> > >
>> > >
>> >
>> >
>>
>>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Fri Aug 28 15:52:45 2020

John,

Just an update, I compared the results using V9.1_beta2 and V9.1, as
you
expected, they are the same.

Thank you for your help! Have a great weekend.
Binyu

On Thu, Aug 27, 2020 at 6:34 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Should you rerun with met-9.1? Generally, the beta releases are
meant for
> testing new development, not production runs. I don't know of any
results
> off the top of my head that would be different between beta2 and
9.1, but
> there certainly could be something. If you'd like to use beta2, it'd
be
> prudent to carefully review the release notes and consider which
changes
> may or may not impact your results. Or you could rerun everything
with
> met-9.1. Or you could just rerun a couple of example cases, diff
your
> results with beta2 output, and decide what you'd like to do from
there.
>
> Listed below are the steps I ran to reorder your data. But I
definitely
> don't recommend this as a general purpose way of doing this.
Instead, it's
> just a hack I used for testing, and it was only for U/V obs, not
other obs
> in the file.
>
> # Run Rscript to dump existing NetCDF file to ascii (probably won't
work on
> wcoss because the ncdf4 library probably doesn't exist)
>
> Rscript
>
> /Volumes/d1/projects/MET/MET_development/MET-
develop/met/scripts/Rscripts/pntnc2ascii.R
> asc2nc_PANO_20200203t12Z.nc > asc2nc_PANO_20200203t12Z.txt
>
>
> # Subset the UGRD and VGRD obs into separate files
>
> grep UGRD asc2nc_PANO_20200203t12Z.txt >
asc2nc_PANO_20200203t12Z_UGRD.txt
>
> grep VGRD asc2nc_PANO_20200203t12Z.txt >
asc2nc_PANO_20200203t12Z_VGRD.txt
>
>
> # Run the paste command to stitch them back together
>
> paste -d
> "\n" asc2nc_PANO_20200203t12Z_UGRD.txt
asc2nc_PANO_20200203t12Z_VGRD.txt
> > asc2nc_PANO_20200203t12Z_sorted.txt
>
>
> # Run ascii2nc to regenerate NetCDF output
>
> ascii2nc asc2nc_PANO_20200203t12Z_sorted.txt
asc2nc_PANO_20200203t12Z_
> sorted.nc
>
> On Thu, Aug 27, 2020 at 3:49 PM binyu.wang at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >
> > John,
> > Thank you. More questions here:
> > 1. I finished running the verification using 9.1_beta2. Should I
rerun
> them
> > using 9.1? Will there be any difference for the point_stat output
> > between using the different versions?
> >
> > 2. How did you do the reorder?  Here is the script I used to do
asc2nc, I
> > used the default config file.
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/asc2grid_uwyo_SKTQ.sh
> >
> > On Thu, Aug 27, 2020 at 5:00 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Binyu,
> > >
> > > Yes, I was just writing up this response...
> > >
> > > OK, I logged onto WCOSS and was able to replicate the behavior
you
> > > describe.
> > >
> > > Here's what I found...
> > >
> > > met-9.1_beta2 does run with this data but met-9.1 does not. But
that's
> > > because there is a bug in met-9.1_beta2 that was fixed in met-
9.1.
> > >
> > > Take a look at this line in point_stat.cc:
> > >
> > >
> >
>
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861
> > >
> > > The bug in beta2 caused this logic not be enforced because
"is_vgrd"
> > never
> > > evaluated to true.
> > >
> > > But the real point here is that by convention, MET expects that
when
> > > processing winds as vectors, each point observation of UGRD is
> > immediately
> > > followed by a VGRD observation at the same level. We enforce
this
> > > requirement to avoid the need of wasting processing time
searching
> > through
> > > all the U/V pairs to match them up. In the data you passed to
> point_stat
> > > your have 175 observations of UGRD followed by 175 observations
of
> VGRD.
> > It
> > > just so happens that the order of the levels in both is
identical...
> but
> > > met-9.1_beta2 didn't actually check that to be sure! So if you
made a
> > > mistake in that regard, met-9.1_beta2 would not have caught it.
> > >
> > > To test things out, I reordered your data to interleave UGRD and
VGRD
> for
> > > each level. And the resulting file ran fine through point_stat.
I've
> > > attached that NetCDF file here.
> > >
> > > Hope that helps clarify.
> > >
> > > Thanks,
> > > John
> > >
> > >
> > > On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> > > >
> > > > Hello John,
> > > >
> > > > Just want to remind that I am still waiting for the answer. I
am
> sorry
> > to
> > > > keep asking questions: I guess I am the one with most of the
> questions.
> > > >
> > > >
> > > >
> > > > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > > Binyu,
> > > > >
> > > > > I believe that WCOSS is down this week and unavailable... at
least
> on
> > > the
> > > > > development side.
> > > > >
> > > > > If you're able to grab that sample data and post it on the
RAL
> > > anonymous
> > > > > ftp site, I could grab it from there for testing.
> > > > >
> > > > > If that possible, please follow the instructions posted
here:
> > > > >
> > > > >
> > > >
> > >
> >
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> > > > >
> > > > > Thanks,
> > > > > John Halley Gotway
> > > > >
> > > > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via RT <
> > > > > met_help at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> > > > > >
> > > > > > Hello again,
> > > > > >
> > > > > > Following your suggestion, I tried to look for more sites
for
> > > > > verification
> > > > > > by using prepbufr data. Unfortunately I always got "0
matching
> > pair"
> > > > > when I
> > > > > > applied PB2NC and ensemble_stat. Here is the directory:
> > > > > >
> > > > > > cd
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> > > > > >
> > > > > > % ./test_bufr.ecf >& bufr.log3
> > > > > >
> > > > > >
> > > > > >  I tried to diagnose the problem by converting the
prepbufr data
> to
> > > > grid
> > > > > > data using point_to_grid, the script is:
> > > > > >
> /gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> > > > > >
> > > > > > One of the output is:
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > > > > > regrid.prebufr.nc.txt
> > > > > >
> > > > > > It seems the PRES field does not look correct to me: it is
over
> > > > ~20000mb,
> > > > > > which is impossible. I confirmed with prepbufr people, the
unit
> > > inside
> > > > of
> > > > > > the prepbufr is mb, and the magnitude should be <1000mb.
But I
> can
> > > not
> > > > > find
> > > > > > where the error is. Any suggestions?
> > > > > >
> > > > > > Thank you.
> > > > > > Binyu
> > > > > >
> > > > > >
> > > > > >
> > > > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > > To keep these issues straight, I just created a separate
RT
> > ticket
> > > > for
> > > > > > > Tatiana to look at. Hopefully you received a copy of it.
> > > > > > >
> > > > > > > Thanks,
> > > > > > > John
> > > > > > >
> > > > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway <
> > > johnhg at ucar.edu>
> > > > > > > wrote:
> > > > > > >
> > > > > > > > Binyu,
> > > > > > > >
> > > > > > > > I'm guessing that the output is not actually the same.
In the
> > > VL1L2
> > > > > and
> > > > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH columns
should now
> > > > > indicate
> > > > > > > that
> > > > > > > > you're using different thresholds: '>=1.0' should now
be
> > changed
> > > to
> > > > > > > > '<=-1.0||>=1.0'
> > > > > > > >
> > > > > > > > You should confirm that those have changed to be
confident
> that
> > > you
> > > > > > > > actually ran the modified configuration. Once you've
> confirmed
> > > > that,
> > > > > if
> > > > > > > the
> > > > > > > > counts and stats are still the same, that's really
just a
> > > function
> > > > of
> > > > > > > your
> > > > > > > > data... about which I have no idea.
> > > > > > > >
> > > > > > > > John
> > > > > > > >
> > > > > > > > On Fri, Jul 31, 2020 at 4:45 PM binyu.wang at noaa.gov
via RT <
> > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > >
> > > > > > > >>
> > > > > > > >> <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > >
> > > > > > > >>
> > > > > > > >> John,
> > > > > > > >>
> > > > > > > >> Should I send another ticket to "Tatiana" about my
log error
> > to
> > > > > upload
> > > > > > > >> point_stat*stat files to METViwer?
> > > > > > > >>
> > > > > > > >> I just tried the new config file (cat_thresh = [
> > <=-1.0||>=1.0,
> > > > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as
using the
> old
> > > > > config
> > > > > > > >> file (
> > > > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I have
too few
> > > > > matching
> > > > > > > >> pairs
> > > > > > > >> (only one or two pairs)?
> > > > > > > >>
> > > > > > > >> Binyu
> > > > > > > >>
> > > > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley Gotway
via RT <
> > > > > > > >> met_help at ucar.edu>
> > > > > > > >> wrote:
> > > > > > > >>
> > > > > > > >> > Binyu,
> > > > > > > >> >
> > > > > > > >> > You're right, the cat_thresh for UGRD of >=1.0
would only
> > > > > threshold
> > > > > > > the
> > > > > > > >> > positive values. If you want to apply both positive
and
> > > negative
> > > > > you
> > > > > > > >> could
> > > > > > > >> > use:
> > > > > > > >> >
> > > > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> > > > > > > >> >
> > > > > > > >> > Table 7.19 in the MET User's Guide describes the
contents
> of
> > > the
> > > > > > VCNT
> > > > > > > >> line
> > > > > > > >> > type:
> > > > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47 PM.png]
> > > > > > > >> >
> > > > > > > >> > Looking at the description of those statistics, it
looks
> > like
> > > > you
> > > > > > may
> > > > > > > be
> > > > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR, and
> > DIR_ABSERR.
> > > > And
> > > > > > > >> > METviewer should be able to plot those statistics
from the
> > > VCNT
> > > > > line
> > > > > > > >> type.
> > > > > > > >> > That being said, I have never actually done this
myself
> for
> > a
> > > > > model
> > > > > > > >> > evaluation.
> > > > > > > >> >
> > > > > > > >> > If METviewer isn't loading VCNT lines or plotting
the
> > > statistics
> > > > > > from
> > > > > > > >> them,
> > > > > > > >> > we'll need to let Tatiana Burek know so that she
can fix
> it.
> > > > > > > >> >
> > > > > > > >> > Thanks,
> > > > > > > >> > John
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM binyu.wang at noaa.gov
via
> RT
> > <
> > > > > > > >> > met_help at ucar.edu> wrote:
> > > > > > > >> >
> > > > > > > >> > >
> > > > > > > >> > > <URL:
> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > >
> > > > > > > >> > >
> > > > > > > >> > > John,
> > > > > > > >> > >
> > > > > > > >> > > That makes sense for wind SPEED. Please look at
the
> > > point_stat
> > > > > > > config:
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> >
> > > > > > > >>
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> > > > > > > >> > > PointStatConfig
> > > > > > > >> > >
> > > > > > > >> > > field = [
> > > > > > > >> > >
> > > > > > > >> > >      {
> > > > > > > >> > >
> > > > > > > >> > >         name       = "UGRD";
> > > > > > > >> > >
> > > > > > > >> > >         level      = [ "P250" ];
> > > > > > > >> > >
> > > > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> > > > > > > >> > >
> > > > > > > >> > >       },
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > the cat_thresh ONLY has positive too, does that
mean
> both
> > > > > > directions
> > > > > > > >> will
> > > > > > > >> > > be considered? (>= refers either ge 5 or le -5)?
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > One more question here: following your
suggestion, I did
> > > > create
> > > > > > > VL1L2
> > > > > > > >> and
> > > > > > > >> > > VCNT,but I got strange error (
> > > > > > > >> > >
> > > > >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> > > > > > > >> when I
> > > > > > > >> > > tried to load to METviewer ( I mentioned that in
one of
> my
> > > > > > previous
> > > > > > > >> > email,
> > > > > > > >> > > but didn't get reply). Is there any plot in
METviwer
> that
> > > can
> > > > > > > compare
> > > > > > > >> the
> > > > > > > >> > > wind direction b/t obs. and forecast?
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > Thank you.
> > > > > > > >> > >
> > > > > > > >> > > Binyu
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley
Gotway via
> RT
> > <
> > > > > > > >> > > met_help at ucar.edu>
> > > > > > > >> > > wrote:
> > > > > > > >> > >
> > > > > > > >> > > > Hi Binyu,
> > > > > > > >> > > >
> > > > > > > >> > > > The wind_thresh is a filtering threshold which
is
> > applied
> > > to
> > > > > the
> > > > > > > >> wind
> > > > > > > >> > > > SPEED. While the individual U and V components
can be
> > > > positive
> > > > > > or
> > > > > > > >> > > negative,
> > > > > > > >> > > > the wind speed is always >= 0.
> > > > > > > >> > > >
> > > > > > > >> > > > The wind_thresh applies to the processing of
the
> VL1L2,
> > > > > VAL1L2,
> > > > > > > and
> > > > > > > >> > VCNT
> > > > > > > >> > > > line types which process U/V pairs. Let's say
you've
> > set:
> > > > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> > > > > > > >> > > >
> > > > > > > >> > > > Instead of seeing only 1 VL1L2 output line,
you'll get
> > > 4...
> > > > > one
> > > > > > > for
> > > > > > > >> > each
> > > > > > > >> > > > wind_thresh chosen.
> > > > > > > >> > > > The first one (wind_thresh = NA) will include
all U/V
> > > pairs
> > > > in
> > > > > > the
> > > > > > > >> > > masking
> > > > > > > >> > > > area.
> > > > > > > >> > > > The second one (wind_thresh ge1) will only
include
> those
> > > U/V
> > > > > > pairs
> > > > > > > >> > whose
> > > > > > > >> > > > corresponding wind speed is at least 1 m/s.
> > > > > > > >> > > > The third and fourth ones (wind_thresh ge5 and
ge10)
> > will
> > > > only
> > > > > > > >> include
> > > > > > > >> > > > those U/V pairs whose corresponding wind speed
is at
> > > least 5
> > > > > m/s
> > > > > > > or
> > > > > > > >> > > 10m/s,
> > > > > > > >> > > > respectively.
> > > > > > > >> > > >
> > > > > > > >> > > > The idea here is checking to see how the model
> performs
> > > > light
> > > > > > and
> > > > > > > >> > > variable
> > > > > > > >> > > > winds (i.e. low wind speed) versus well-defined
winds
> > > (i.e.
> > > > > high
> > > > > > > >> wind
> > > > > > > >> > > > speed).
> > > > > > > >> > > >
> > > > > > > >> > > > Thanks,
> > > > > > > >> > > > John
> > > > > > > >> > > >
> > > > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM
binyu.wang at noaa.gov
> via
> > > RT
> > > > <
> > > > > > > >> > > > met_help at ucar.edu> wrote:
> > > > > > > >> > > >
> > > > > > > >> > > > >
> > > > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074 was
acted
> > upon.
> > > > > > > >> > > > > Transaction: Ticket created by
binyu.wang at noaa.gov
> > > > > > > >> > > > >        Queue: met_help
> > > > > > > >> > > > >      Subject: wind_thresh
> > > > > > > >> > > > >        Owner: Nobody
> > > > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> > > > > > > >> > > > >       Status: new
> > > > > > > >> > > > >  Ticket <URL:
> > > > > > > >> >
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> > > > > > > >> > > >
> > > > > > > >> > > > >
> > > > > > > >> > > > >
> > > > > > > >> > > > > Hello,
> > > > > > > >> > > > >
> > > > > > > >> > > > > I have a question about setting up
wind_thresh in
> > config
> > > > > file:
> > > > > > > >> > because
> > > > > > > >> > > > the
> > > > > > > >> > > > > wind can be negative and positive, so how to
set
> them
> > > > > > correctly?
> > > > > > > >> > > > > e.g, which is correct:
> > > > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
> >=5.0&&<=10
> > > > > > > >> ||>=-10&&<=-5,
> > > > > > > >> > > > > >=10.0||<=-10 ];
> > > > > > > >> > > > > or
> > > > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,  >=5.0&&<=10,
>=20  ];
> > > > > > > >> > > > >
> > > > > > > >> > > > > Thank you.
> > > > > > > >> > > > > Binyu
> > > > > > > >> > > > >
> > > > > > > >> > > > >
> > > > > > > >> > > >
> > > > > > > >> > > >
> > > > > > > >> > >
> > > > > > > >> > >
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >>
> > > > > > > >>
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Fri Aug 28 15:58:42 2020

Hello John,

I am sorry that some inputs are gone.

I did do a quick test as you suggested, and it seems there was no
problem,
here is the script and log file:
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/point_stat_burf.sh
>point_stat_burf.sh.log

However, when I tried my ensemble_stat, I got same error as before,
the log
file is:
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf/bufr.log4

All the inputs are on WCOSS now. Thank you.
Binyu

On Fri, Aug 28, 2020 at 4:28 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Binyu,
>
> OK I'm taking a look in this log file:
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf/bufr.log3
>
> Whenever folks get 0 matched pairs from Point-Stat, I always advise
that
> they increase the verbosity level to list out counts of reason codes
> explaining why obs were or were not used, such as below:
>
> DEBUG 3: Number of matched pairs   = 132
>
> DEBUG 3: Observations processed    = 89893
>
> DEBUG 3: Rejected: station id      = 0
>
> DEBUG 3: Rejected: obs type        = 79788
>
> DEBUG 3: Rejected: valid time      = 0
>
> DEBUG 3: Rejected: bad obs value   = 0
>
> DEBUG 3: Rejected: off the grid    = 5
>
> DEBUG 3: Rejected: topography      = 0
>
> DEBUG 3: Rejected: level mismatch  = 9219
>
> DEBUG 3: Rejected: quality marker  = 0
>
> DEBUG 3: Rejected: message type    = 344
>
> DEBUG 3: Rejected: masking region  = 405
>
> DEBUG 3: Rejected: bad fcst value  = 0
>
> DEBUG 3: Rejected: bad climo mean  = 0
>
> DEBUG 3: Rejected: bad climo stdev = 0
>
> DEBUG 3: Rejected: duplicates      = 0
>
>
> Unfortunately though, Ensemble-Stat does NOT track and report the
same
> counts!
>
>
> So I wanted to actually run Point-Stat using 1 of the ensemble
members with
> those same point observations at verbosity level 3.
>
>
> BUT when I logged on WCOSS, I could not actually locate those files.
So you
> could either...
>
> (1) Do this testing yourself. Run Point-Stat to compare 1 ensemble
member
> to these obs and check the verbosity level 3 output to see why there
are 0
> matched pairs.
>
> (2) Or you could post the following files to our anonymous FTP site
> following these instructions:
>
>
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
>
>
> I'd need:
>
> (1) Ens file:
>
> gefs01.t00z.grid3.f12.grib2
>
> (2) Obs file:
>
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/
> gdas.2020020300.prebufr.f00.nc
>
> (3) Ensemble-Stat config file:
>
>
>
/gpfs/dell1/ptmp/Binyu.Wang/gens/tmpnwprd/verf_g2o_gefs_00_dev.25142/verf_g2o_ens_gens_config.bufr
>
>
> Please just let me know.
>
>
> Thanks,
> John
>
> On Thu, Aug 27, 2020 at 5:26 PM John Halley Gotway <johnhg at ucar.edu>
> wrote:
>
> > Sorry, I'm headed out for the day. But will keep this message
marked
> > unread and get back to it when I get a chance.
> >
> > Thanks,
> > John
> >
> > On Thu, Aug 27, 2020 at 5:00 PM binyu.wang at noaa.gov via RT <
> > met_help at ucar.edu> wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >>
> >> Got it, thank you very much!
> >> Before you are very tired of answering my questions, there is one
more
> >> from
> >> two weeks ago:
> >>
> >>  I tried to do gefs verification by using prepbufr data.
Unfortunately I
> >> always got "0 matching pair" when I applied PB2NC and
ensemble_stat.
> Here
> >> is the directory:
> >>
> >> cd
> >>
> >>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> >>
> >> % ./test_bufr.ecf >& bufr.log3
> >>
> >>
> >>  I tried to diagnose the problem by converting the prepbufr data
to grid
> >> data using point_to_grid, the script is:
> >>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> >>
> >> One of the output is:
> >>
> >>
> >>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> >> regrid.prebufr.nc.txt
> >>
> >> It seems the PRES field does not look correct to me: it is over
> ~20000mb,
> >> which is impossible. I confirmed with prepbufr people, the unit
inside
> of
> >> the prepbufr is mb, and the magnitude should be <1000mb. But I
can not
> >> find
> >> where the error is. Any suggestions?
> >>
> >>
> >> John Halley Gotway via RT <met_help at ucar.edu>
> >>
> >> On Thu, Aug 27, 2020 at 6:34 PM John Halley Gotway via RT <
> >> met_help at ucar.edu>
> >> wrote:
> >>
> >> > Should you rerun with met-9.1? Generally, the beta releases are
meant
> >> for
> >> > testing new development, not production runs. I don't know of
any
> >> results
> >> > off the top of my head that would be different between beta2
and 9.1,
> >> but
> >> > there certainly could be something. If you'd like to use beta2,
it'd
> be
> >> > prudent to carefully review the release notes and consider
which
> changes
> >> > may or may not impact your results. Or you could rerun
everything with
> >> > met-9.1. Or you could just rerun a couple of example cases,
diff your
> >> > results with beta2 output, and decide what you'd like to do
from
> there.
> >> >
> >> > Listed below are the steps I ran to reorder your data. But I
> definitely
> >> > don't recommend this as a general purpose way of doing this.
Instead,
> >> it's
> >> > just a hack I used for testing, and it was only for U/V obs,
not other
> >> obs
> >> > in the file.
> >> >
> >> > # Run Rscript to dump existing NetCDF file to ascii (probably
won't
> >> work on
> >> > wcoss because the ncdf4 library probably doesn't exist)
> >> >
> >> > Rscript
> >> >
> >> >
> >>
> /Volumes/d1/projects/MET/MET_development/MET-
develop/met/scripts/Rscripts/pntnc2ascii.R
> >> > asc2nc_PANO_20200203t12Z.nc > asc2nc_PANO_20200203t12Z.txt
> >> >
> >> >
> >> > # Subset the UGRD and VGRD obs into separate files
> >> >
> >> > grep UGRD asc2nc_PANO_20200203t12Z.txt >
> >> asc2nc_PANO_20200203t12Z_UGRD.txt
> >> >
> >> > grep VGRD asc2nc_PANO_20200203t12Z.txt >
> >> asc2nc_PANO_20200203t12Z_VGRD.txt
> >> >
> >> >
> >> > # Run the paste command to stitch them back together
> >> >
> >> > paste -d
> >> > "\n" asc2nc_PANO_20200203t12Z_UGRD.txt
> asc2nc_PANO_20200203t12Z_VGRD.txt
> >> > > asc2nc_PANO_20200203t12Z_sorted.txt
> >> >
> >> >
> >> > # Run ascii2nc to regenerate NetCDF output
> >> >
> >> > ascii2nc asc2nc_PANO_20200203t12Z_sorted.txt
asc2nc_PANO_20200203t12Z_
> >> > sorted.nc
> >> >
> >> > On Thu, Aug 27, 2020 at 3:49 PM binyu.wang at noaa.gov via RT <
> >> > met_help at ucar.edu> wrote:
> >> >
> >> > >
> >> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
>
> >> > >
> >> > > John,
> >> > > Thank you. More questions here:
> >> > > 1. I finished running the verification using 9.1_beta2.
Should I
> rerun
> >> > them
> >> > > using 9.1? Will there be any difference for the point_stat
output
> >> > > between using the different versions?
> >> > >
> >> > > 2. How did you do the reorder?  Here is the script I used to
do
> >> asc2nc, I
> >> > > used the default config file.
> >> > >
> >> > >
> >> >
> >>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/asc2grid_uwyo_SKTQ.sh
> >> > >
> >> > > On Thu, Aug 27, 2020 at 5:00 PM John Halley Gotway via RT <
> >> > > met_help at ucar.edu>
> >> > > wrote:
> >> > >
> >> > > > Binyu,
> >> > > >
> >> > > > Yes, I was just writing up this response...
> >> > > >
> >> > > > OK, I logged onto WCOSS and was able to replicate the
behavior you
> >> > > > describe.
> >> > > >
> >> > > > Here's what I found...
> >> > > >
> >> > > > met-9.1_beta2 does run with this data but met-9.1 does not.
But
> >> that's
> >> > > > because there is a bug in met-9.1_beta2 that was fixed in
met-9.1.
> >> > > >
> >> > > > Take a look at this line in point_stat.cc:
> >> > > >
> >> > > >
> >> > >
> >> >
> >>
>
https://github.com/dtcenter/MET/blob/2c9ae440a84024fbf62caa64c1b747f9a912236f/met/src/tools/core/point_stat/point_stat.cc#L861
> >> > > >
> >> > > > The bug in beta2 caused this logic not be enforced because
> "is_vgrd"
> >> > > never
> >> > > > evaluated to true.
> >> > > >
> >> > > > But the real point here is that by convention, MET expects
that
> when
> >> > > > processing winds as vectors, each point observation of UGRD
is
> >> > > immediately
> >> > > > followed by a VGRD observation at the same level. We
enforce this
> >> > > > requirement to avoid the need of wasting processing time
searching
> >> > > through
> >> > > > all the U/V pairs to match them up. In the data you passed
to
> >> > point_stat
> >> > > > your have 175 observations of UGRD followed by 175
observations of
> >> > VGRD.
> >> > > It
> >> > > > just so happens that the order of the levels in both is
> identical...
> >> > but
> >> > > > met-9.1_beta2 didn't actually check that to be sure! So if
you
> made
> >> a
> >> > > > mistake in that regard, met-9.1_beta2 would not have caught
it.
> >> > > >
> >> > > > To test things out, I reordered your data to interleave
UGRD and
> >> VGRD
> >> > for
> >> > > > each level. And the resulting file ran fine through
point_stat.
> I've
> >> > > > attached that NetCDF file here.
> >> > > >
> >> > > > Hope that helps clarify.
> >> > > >
> >> > > > Thanks,
> >> > > > John
> >> > > >
> >> > > >
> >> > > > On Thu, Aug 27, 2020 at 2:28 PM binyu.wang at noaa.gov via RT
<
> >> > > > met_help at ucar.edu> wrote:
> >> > > >
> >> > > > >
> >> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
> >> > > > >
> >> > > > > Hello John,
> >> > > > >
> >> > > > > Just want to remind that I am still waiting for the
answer. I am
> >> > sorry
> >> > > to
> >> > > > > keep asking questions: I guess I am the one with most of
the
> >> > questions.
> >> > > > >
> >> > > > >
> >> > > > >
> >> > > > > On Tue, Aug 11, 2020 at 6:36 PM John Halley Gotway via RT
<
> >> > > > > met_help at ucar.edu>
> >> > > > > wrote:
> >> > > > >
> >> > > > > > Binyu,
> >> > > > > >
> >> > > > > > I believe that WCOSS is down this week and
unavailable... at
> >> least
> >> > on
> >> > > > the
> >> > > > > > development side.
> >> > > > > >
> >> > > > > > If you're able to grab that sample data and post it on
the RAL
> >> > > > anonymous
> >> > > > > > ftp site, I could grab it from there for testing.
> >> > > > > >
> >> > > > > > If that possible, please follow the instructions posted
here:
> >> > > > > >
> >> > > > > >
> >> > > > >
> >> > > >
> >> > >
> >> >
> >>
> http://dtcenter.org/community-code/model-evaluation-tools-met/met-
help-desk#ftp
> >> > > > > >
> >> > > > > > Thanks,
> >> > > > > > John Halley Gotway
> >> > > > > >
> >> > > > > > On Thu, Aug 6, 2020 at 1:30 PM binyu.wang at noaa.gov via
RT <
> >> > > > > > met_help at ucar.edu>
> >> > > > > > wrote:
> >> > > > > >
> >> > > > > > >
> >> > > > > > > <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >> >
> >> > > > > > >
> >> > > > > > > Hello again,
> >> > > > > > >
> >> > > > > > > Following your suggestion, I tried to look for more
sites
> for
> >> > > > > > verification
> >> > > > > > > by using prepbufr data. Unfortunately I always got "0
> matching
> >> > > pair"
> >> > > > > > when I
> >> > > > > > > applied PB2NC and ensemble_stat. Here is the
directory:
> >> > > > > > >
> >> > > > > > > cd
> >> > > > > > >
> >> > > > > > >
> >> > > > > >
> >> > > > >
> >> > > >
> >> > >
> >> >
> >>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf
> >> > > > > > >
> >> > > > > > > % ./test_bufr.ecf >& bufr.log3
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >  I tried to diagnose the problem by converting the
prepbufr
> >> data
> >> > to
> >> > > > > grid
> >> > > > > > > data using point_to_grid, the script is:
> >> > > > > > >
> >> >
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/Scripts/MET/pt2grid.sh
> >> > > > > > >
> >> > > > > > > One of the output is:
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > >
> >> > > > >
> >> > > >
> >> > >
> >> >
> >>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> >> > > > > > > regrid.prebufr.nc.txt
> >> > > > > > >
> >> > > > > > > It seems the PRES field does not look correct to me:
it is
> >> over
> >> > > > > ~20000mb,
> >> > > > > > > which is impossible. I confirmed with prepbufr
people, the
> >> unit
> >> > > > inside
> >> > > > > of
> >> > > > > > > the prepbufr is mb, and the magnitude should be
<1000mb.
> But I
> >> > can
> >> > > > not
> >> > > > > > find
> >> > > > > > > where the error is. Any suggestions?
> >> > > > > > >
> >> > > > > > > Thank you.
> >> > > > > > > Binyu
> >> > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > > > On Fri, Jul 31, 2020 at 6:58 PM John Halley Gotway
via RT <
> >> > > > > > > met_help at ucar.edu>
> >> > > > > > > wrote:
> >> > > > > > >
> >> > > > > > > > To keep these issues straight, I just created a
separate
> RT
> >> > > ticket
> >> > > > > for
> >> > > > > > > > Tatiana to look at. Hopefully you received a copy
of it.
> >> > > > > > > >
> >> > > > > > > > Thanks,
> >> > > > > > > > John
> >> > > > > > > >
> >> > > > > > > > On Fri, Jul 31, 2020 at 4:52 PM John Halley Gotway
<
> >> > > > johnhg at ucar.edu>
> >> > > > > > > > wrote:
> >> > > > > > > >
> >> > > > > > > > > Binyu,
> >> > > > > > > > >
> >> > > > > > > > > I'm guessing that the output is not actually the
same.
> In
> >> the
> >> > > > VL1L2
> >> > > > > > and
> >> > > > > > > > > VCNT lines, the FCST_THRESH and OBS_THRESH
columns
> should
> >> now
> >> > > > > > indicate
> >> > > > > > > > that
> >> > > > > > > > > you're using different thresholds: '>=1.0' should
now be
> >> > > changed
> >> > > > to
> >> > > > > > > > > '<=-1.0||>=1.0'
> >> > > > > > > > >
> >> > > > > > > > > You should confirm that those have changed to be
> confident
> >> > that
> >> > > > you
> >> > > > > > > > > actually ran the modified configuration. Once
you've
> >> > confirmed
> >> > > > > that,
> >> > > > > > if
> >> > > > > > > > the
> >> > > > > > > > > counts and stats are still the same, that's
really just
> a
> >> > > > function
> >> > > > > of
> >> > > > > > > > your
> >> > > > > > > > > data... about which I have no idea.
> >> > > > > > > > >
> >> > > > > > > > > John
> >> > > > > > > > >
> >> > > > > > > > > On Fri, Jul 31, 2020 at 4:45 PM
binyu.wang at noaa.gov via
> >> RT <
> >> > > > > > > > > met_help at ucar.edu> wrote:
> >> > > > > > > > >
> >> > > > > > > > >>
> >> > > > > > > > >> <URL:
> >> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >> > > >
> >> > > > > > > > >>
> >> > > > > > > > >> John,
> >> > > > > > > > >>
> >> > > > > > > > >> Should I send another ticket to "Tatiana" about
my log
> >> error
> >> > > to
> >> > > > > > upload
> >> > > > > > > > >> point_stat*stat files to METViwer?
> >> > > > > > > > >>
> >> > > > > > > > >> I just tried the new config file (cat_thresh = [
> >> > > <=-1.0||>=1.0,
> >> > > > > > > > >> <=-5.0||>=5.0 ];) and the output are the SAME as
using
> >> the
> >> > old
> >> > > > > > config
> >> > > > > > > > >> file (
> >> > > > > > > > >>  cat_thresh = [ >=1.0, >=5.0 ];). Is this b/c I
have
> too
> >> few
> >> > > > > > matching
> >> > > > > > > > >> pairs
> >> > > > > > > > >> (only one or two pairs)?
> >> > > > > > > > >>
> >> > > > > > > > >> Binyu
> >> > > > > > > > >>
> >> > > > > > > > >> On Fri, Jul 31, 2020 at 6:10 PM John Halley
Gotway via
> >> RT <
> >> > > > > > > > >> met_help at ucar.edu>
> >> > > > > > > > >> wrote:
> >> > > > > > > > >>
> >> > > > > > > > >> > Binyu,
> >> > > > > > > > >> >
> >> > > > > > > > >> > You're right, the cat_thresh for UGRD of >=1.0
would
> >> only
> >> > > > > > threshold
> >> > > > > > > > the
> >> > > > > > > > >> > positive values. If you want to apply both
positive
> and
> >> > > > negative
> >> > > > > > you
> >> > > > > > > > >> could
> >> > > > > > > > >> > use:
> >> > > > > > > > >> >
> >> > > > > > > > >> > cat_thresh = [ <=-1.0||>=1.0, <=-5.0||>=5.0 ];
> >> > > > > > > > >> >
> >> > > > > > > > >> > Table 7.19 in the MET User's Guide describes
the
> >> contents
> >> > of
> >> > > > the
> >> > > > > > > VCNT
> >> > > > > > > > >> line
> >> > > > > > > > >> > type:
> >> > > > > > > > >> > [image: Screen Shot 2020-07-31 at 4.06.47
PM.png]
> >> > > > > > > > >> >
> >> > > > > > > > >> > Looking at the description of those
statistics, it
> >> looks
> >> > > like
> >> > > > > you
> >> > > > > > > may
> >> > > > > > > > be
> >> > > > > > > > >> > interested in FDIR, ORDIR, VDIFF_DIR, DIR_ERR,
and
> >> > > DIR_ABSERR.
> >> > > > > And
> >> > > > > > > > >> > METviewer should be able to plot those
statistics
> from
> >> the
> >> > > > VCNT
> >> > > > > > line
> >> > > > > > > > >> type.
> >> > > > > > > > >> > That being said, I have never actually done
this
> myself
> >> > for
> >> > > a
> >> > > > > > model
> >> > > > > > > > >> > evaluation.
> >> > > > > > > > >> >
> >> > > > > > > > >> > If METviewer isn't loading VCNT lines or
plotting the
> >> > > > statistics
> >> > > > > > > from
> >> > > > > > > > >> them,
> >> > > > > > > > >> > we'll need to let Tatiana Burek know so that
she can
> >> fix
> >> > it.
> >> > > > > > > > >> >
> >> > > > > > > > >> > Thanks,
> >> > > > > > > > >> > John
> >> > > > > > > > >> >
> >> > > > > > > > >> >
> >> > > > > > > > >> >
> >> > > > > > > > >> > On Fri, Jul 31, 2020 at 3:34 PM
binyu.wang at noaa.gov
> >> via
> >> > RT
> >> > > <
> >> > > > > > > > >> > met_help at ucar.edu> wrote:
> >> > > > > > > > >> >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > <URL:
> >> > > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >> > > > > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > John,
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > That makes sense for wind SPEED. Please look
at the
> >> > > > point_stat
> >> > > > > > > > config:
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> >
> >> > > > > > > > >>
> >> > > > > > > >
> >> > > > > > >
> >> > > > > >
> >> > > > >
> >> > > >
> >> > >
> >> >
> >>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_gec/parm/
> >> > > > > > > > >> > > PointStatConfig
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > field = [
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >      {
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >         name       = "UGRD";
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >         level      = [ "P250" ];
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >         cat_thresh = [ >=1.0, >=5.0 ];
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >       },
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > the cat_thresh ONLY has positive too, does
that
> mean
> >> > both
> >> > > > > > > directions
> >> > > > > > > > >> will
> >> > > > > > > > >> > > be considered? (>= refers either ge 5 or le
-5)?
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > One more question here: following your
suggestion,
> I
> >> did
> >> > > > > create
> >> > > > > > > > VL1L2
> >> > > > > > > > >> and
> >> > > > > > > > >> > > VCNT,but I got strange error (
> >> > > > > > > > >> > >
> >> > > > > >
> >>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/METviewer_AWS/g2g)
> >> > > > > > > > >> when I
> >> > > > > > > > >> > > tried to load to METviewer ( I mentioned
that in
> one
> >> of
> >> > my
> >> > > > > > > previous
> >> > > > > > > > >> > email,
> >> > > > > > > > >> > > but didn't get reply). Is there any plot in
> METviwer
> >> > that
> >> > > > can
> >> > > > > > > > compare
> >> > > > > > > > >> the
> >> > > > > > > > >> > > wind direction b/t obs. and forecast?
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > Thank you.
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > Binyu
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > On Fri, Jul 31, 2020 at 4:07 PM John Halley
Gotway
> >> via
> >> > RT
> >> > > <
> >> > > > > > > > >> > > met_help at ucar.edu>
> >> > > > > > > > >> > > wrote:
> >> > > > > > > > >> > >
> >> > > > > > > > >> > > > Hi Binyu,
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > The wind_thresh is a filtering threshold
which is
> >> > > applied
> >> > > > to
> >> > > > > > the
> >> > > > > > > > >> wind
> >> > > > > > > > >> > > > SPEED. While the individual U and V
components
> can
> >> be
> >> > > > > positive
> >> > > > > > > or
> >> > > > > > > > >> > > negative,
> >> > > > > > > > >> > > > the wind speed is always >= 0.
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > The wind_thresh applies to the processing
of the
> >> > VL1L2,
> >> > > > > > VAL1L2,
> >> > > > > > > > and
> >> > > > > > > > >> > VCNT
> >> > > > > > > > >> > > > line types which process U/V pairs. Let's
say
> >> you've
> >> > > set:
> >> > > > > > > > >> > > > wind_thresh = [ NA, ge1, ge5, ge10 ];
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > Instead of seeing only 1 VL1L2 output
line,
> you'll
> >> get
> >> > > > 4...
> >> > > > > > one
> >> > > > > > > > for
> >> > > > > > > > >> > each
> >> > > > > > > > >> > > > wind_thresh chosen.
> >> > > > > > > > >> > > > The first one (wind_thresh = NA) will
include all
> >> U/V
> >> > > > pairs
> >> > > > > in
> >> > > > > > > the
> >> > > > > > > > >> > > masking
> >> > > > > > > > >> > > > area.
> >> > > > > > > > >> > > > The second one (wind_thresh ge1) will only
> include
> >> > those
> >> > > > U/V
> >> > > > > > > pairs
> >> > > > > > > > >> > whose
> >> > > > > > > > >> > > > corresponding wind speed is at least 1
m/s.
> >> > > > > > > > >> > > > The third and fourth ones (wind_thresh ge5
and
> >> ge10)
> >> > > will
> >> > > > > only
> >> > > > > > > > >> include
> >> > > > > > > > >> > > > those U/V pairs whose corresponding wind
speed is
> >> at
> >> > > > least 5
> >> > > > > > m/s
> >> > > > > > > > or
> >> > > > > > > > >> > > 10m/s,
> >> > > > > > > > >> > > > respectively.
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > The idea here is checking to see how the
model
> >> > performs
> >> > > > > light
> >> > > > > > > and
> >> > > > > > > > >> > > variable
> >> > > > > > > > >> > > > winds (i.e. low wind speed) versus well-
defined
> >> winds
> >> > > > (i.e.
> >> > > > > > high
> >> > > > > > > > >> wind
> >> > > > > > > > >> > > > speed).
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > Thanks,
> >> > > > > > > > >> > > > John
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > On Fri, Jul 31, 2020 at 7:19 AM
> >> binyu.wang at noaa.gov
> >> > via
> >> > > > RT
> >> > > > > <
> >> > > > > > > > >> > > > met_help at ucar.edu> wrote:
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > >
> >> > > > > > > > >> > > > > Fri Jul 31 07:18:27 2020: Request 96074
was
> acted
> >> > > upon.
> >> > > > > > > > >> > > > > Transaction: Ticket created by
> >> binyu.wang at noaa.gov
> >> > > > > > > > >> > > > >        Queue: met_help
> >> > > > > > > > >> > > > >      Subject: wind_thresh
> >> > > > > > > > >> > > > >        Owner: Nobody
> >> > > > > > > > >> > > > >   Requestors: binyu.wang at noaa.gov
> >> > > > > > > > >> > > > >       Status: new
> >> > > > > > > > >> > > > >  Ticket <URL:
> >> > > > > > > > >> >
> >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > > >
> >> > > > > > > > >> > > > >
> >> > > > > > > > >> > > > > Hello,
> >> > > > > > > > >> > > > >
> >> > > > > > > > >> > > > > I have a question about setting up
wind_thresh
> in
> >> > > config
> >> > > > > > file:
> >> > > > > > > > >> > because
> >> > > > > > > > >> > > > the
> >> > > > > > > > >> > > > > wind can be negative and positive, so
how to
> set
> >> > them
> >> > > > > > > correctly?
> >> > > > > > > > >> > > > > e.g, which is correct:
> >> > > > > > > > >> > > > > wind_thresh = [ >=1&&<=5 || >=-5&&<=-1 ,
> >> > >=5.0&&<=10
> >> > > > > > > > >> ||>=-10&&<=-5,
> >> > > > > > > > >> > > > > >=10.0||<=-10 ];
> >> > > > > > > > >> > > > > or
> >> > > > > > > > >> > > > > |wind_thresh| = [ >=1&&<=5,
>=5.0&&<=10,  >=20
> >> ];
> >> > > > > > > > >> > > > >
> >> > > > > > > > >> > > > > Thank you.
> >> > > > > > > > >> > > > > Binyu
> >> > > > > > > > >> > > > >
> >> > > > > > > > >> > > > >
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> > >
> >> > > > > > > > >> >
> >> > > > > > > > >> >
> >> > > > > > > > >>
> >> > > > > > > > >>
> >> > > > > > > >
> >> > > > > > > >
> >> > > > > > >
> >> > > > > > >
> >> > > > > >
> >> > > > > >
> >> > > > >
> >> > > > >
> >> > > >
> >> > > >
> >> > >
> >> > >
> >> >
> >> >
> >>
> >>
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Mon Aug 31 15:15:58 2020

Binyu,

Thanks for the update log files to look investigate.

I'm looking in:
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/point_stat_burf.sh.log*
And in:
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf/bufr.log4*

Looking carefully, I see that you are using 2 different point
observation
files:
In Point-Stat:

*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/gdas.2020020200.prebufr.f00.nc
<http://gdas.2020020200.prebufr.f00.nc>*

In Ensemble-Stat:
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/gdas.2020020300.prebufr.f00.nc
<http://gdas.2020020300.prebufr.f00.nc>*

The Point-Stat one contains *1328047* observations:

*ncdump -h
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/gdas.2020020200.prebufr.f00.nc
<http://gdas.2020020200.prebufr.f00.nc> | grep nobs*

While the Ensemble-Stat one only contains *193664*:

*ncdump -h
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/gdas.2020020300.prebufr.f00.nc
<http://gdas.2020020300.prebufr.f00.nc> | grep nobs*

I'd recommend testing with the same input point observation files
before
looking at anything else.

Thanks,
John

------------------------------------------------
Subject: wind_thresh
From: binyu.wang at noaa.gov
Time: Wed Sep 09 20:59:18 2020

Hello John,

Just let you know that I have the problem solved. Thank you for your
time
and help.
Binyu

On Mon, Aug 31, 2020 at 5:16 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:

> Binyu,
>
> Thanks for the update log files to look investigate.
>
> I'm looking in:
>
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/point_stat_burf.sh.log*
> And in:
>
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf/bufr.log4*
>
> Looking carefully, I see that you are using 2 different point
observation
> files:
> In Point-Stat:
>
>
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> gdas.2020020200.prebufr.f00.nc
> <http://gdas.2020020200.prebufr.f00.nc>*
>
> In Ensemble-Stat:
>
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/
> gdas.2020020300.prebufr.f00.nc
> <http://gdas.2020020300.prebufr.f00.nc>*
>
> The Point-Stat one contains *1328047* observations:
>
> *ncdump -h
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> gdas.2020020200.prebufr.f00.nc
> <http://gdas.2020020200.prebufr.f00.nc> | grep nobs*
>
> While the Ensemble-Stat one only contains *193664*:
>
> *ncdump -h
>
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/
> gdas.2020020300.prebufr.f00.nc
> <http://gdas.2020020300.prebufr.f00.nc> | grep nobs*
>
> I'd recommend testing with the same input point observation files
before
> looking at anything else.
>
> Thanks,
> John
>
>

------------------------------------------------
Subject: wind_thresh
From: John Halley Gotway
Time: Thu Sep 10 12:23:19 2020

Great, glad to hear it. Thanks for confirming.

John

On Wed, Sep 9, 2020 at 8:59 PM binyu.wang at noaa.gov via RT
<met_help at ucar.edu>
wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=96074 >
>
> Hello John,
>
> Just let you know that I have the problem solved. Thank you for your
time
> and help.
> Binyu
>
> On Mon, Aug 31, 2020 at 5:16 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Binyu,
> >
> > Thanks for the update log files to look investigate.
> >
> > I'm looking in:
> >
> >
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/point_stat_burf.sh.log*
> > And in:
> >
> >
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ecf/bufr.log4*
> >
> > Looking carefully, I see that you are using 2 different point
observation
> > files:
> > In Point-Stat:
> >
> >
> >
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > gdas.2020020200.prebufr.f00.nc
> > <http://gdas.2020020200.prebufr.f00.nc>*
> >
> > In Ensemble-Stat:
> >
> >
>
*/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/
> > gdas.2020020300.prebufr.f00.nc
> > <http://gdas.2020020300.prebufr.f00.nc>*
> >
> > The Point-Stat one contains *1328047* observations:
> >
> > *ncdump -h
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/verf_met_ens_g2o/ush/COMOUT/
> > gdas.2020020200.prebufr.f00.nc
> > <http://gdas.2020020200.prebufr.f00.nc> | grep nobs*
> >
> > While the Ensemble-Stat one only contains *193664*:
> >
> > *ncdump -h
> >
> >
>
/gpfs/dell2/emc/modeling/noscrub/Binyu.Wang/MET/grid2grid/com/gefs.20200203/
> > gdas.2020020300.prebufr.f00.nc
> > <http://gdas.2020020300.prebufr.f00.nc> | grep nobs*
> >
> > I'd recommend testing with the same input point observation files
before
> > looking at anything else.
> >
> > Thanks,
> > John
> >
> >
>
>

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