[Met_help] [rt.rap.ucar.edu #72431] History for suggestion for tc_pairs

John Halley Gotway via RT met_help at ucar.edu
Mon Aug 17 12:11:14 MDT 2015


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

Hi there,

Just a quick suggestion for tc_pairs. When using tc_pairs, if there are
less than 18 fields in the ATCF file, tc_pairs will issue the following
warning:

WARNING:
WARNING: bool operator>>(istream & in, ATCFLine & t) -> found fewer than
the expected number of elements (8<18) in ATCF line:
WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
WARNING:

​and then stop reading the file. ​This is somewhat of a pain in that the
majority of the ATCF files I need to verify do not have a full 18
columns, and I need to pre-process them all to ensure tc_pairs does not
complain. Can a flag option be included in the namelist or
configure file to automatically fill out a ATCF file with null values if
the number of columns is below the needed amount? I don't know
how much work this would entail, but it would be a great option to have.
tc_pairs will currently exit if it encounters a non-standard line,
even if it doesn't correspond to one of the models you're trying to verify.

Aside from this rather small issue, I really appreciate all of the
functionality available with tc_pairs and tc_stat. Thanks for all your hard
work!

Take care,
Andy

-- 
Andrew Penny
Hurricane Model Diagnostician
SRG / National Hurricane Center
11691 SW 17th Street
Miami, FL 33165-2149
phone: 305.229.4457
email: andrew.penny at noaa.gov


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

Subject: suggestion for tc_pairs
From: John Halley Gotway
Time: Wed Jun 24 09:30:25 2015

Andy,

Sure, that'd be simple enough.  I'll add it to our list of development
tasks.  Can you tell me though, what is the source of this ATCF data
that
you're using?  Is it real hurricane data or is it some novel use of
the
ATCF file format?  Knowing that would help put your request in
context.  Is
someone out there generating real ATCF data that only has 8 columns?
If
so, perhaps that we should change the minimum required from 18 to 8.

John

On Wed, Jun 24, 2015 at 4:42 AM, Andrew Penny - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:

>
> Wed Jun 24 04:42:49 2015: Request 72431 was acted upon.
> Transaction: Ticket created by andrew.penny at noaa.gov
>        Queue: met_help
>      Subject: suggestion for tc_pairs
>        Owner: Nobody
>   Requestors: andrew.penny at noaa.gov
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
>
>
> Hi there,
>
> Just a quick suggestion for tc_pairs. When using tc_pairs, if there
are
> less than 18 fields in the ATCF file, tc_pairs will issue the
following
> warning:
>
> WARNING:
> WARNING: bool operator>>(istream & in, ATCFLine & t) -> found fewer
than
> the expected number of elements (8<18) in ATCF line:
> WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
> WARNING:
>
> ​and then stop reading the file. ​This is somewhat of a pain in that
the
> majority of the ATCF files I need to verify do not have a full 18
> columns, and I need to pre-process them all to ensure tc_pairs does
not
> complain. Can a flag option be included in the namelist or
> configure file to automatically fill out a ATCF file with null
values if
> the number of columns is below the needed amount? I don't know
> how much work this would entail, but it would be a great option to
have.
> tc_pairs will currently exit if it encounters a non-standard line,
> even if it doesn't correspond to one of the models you're trying to
verify.
>
> Aside from this rather small issue, I really appreciate all of the
> functionality available with tc_pairs and tc_stat. Thanks for all
your hard
> work!
>
> Take care,
> Andy
>
> --
> Andrew Penny
> Hurricane Model Diagnostician
> SRG / National Hurricane Center
> 11691 SW 17th Street
> Miami, FL 33165-2149
> phone: 305.229.4457
> email: andrew.penny at noaa.gov
>
>

------------------------------------------------
Subject: suggestion for tc_pairs
From: Andrew Penny - NOAA Affiliate
Time: Wed Jun 24 10:35:21 2015

Hi John,

I've recently been doing a lot of verification for retrospective runs
(to
evaluate proposed model upgrades). To be honest, I'm not sure why some
ATCF
files are more filled out than others. Perhaps the other fields are
left
out to save time since just the track and intensity are used to verify
the
retrospective runs. In fact, sometimes there isn't a column for
intensity,
as some models are not used for intensity guidance.

Hope that helps. Let me know if you need more information.

Thanks and take care,
Andy

On Wed, Jun 24, 2015 at 11:30 AM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:

> Andy,
>
> Sure, that'd be simple enough.  I'll add it to our list of
development
> tasks.  Can you tell me though, what is the source of this ATCF data
that
> you're using?  Is it real hurricane data or is it some novel use of
the
> ATCF file format?  Knowing that would help put your request in
context.  Is
> someone out there generating real ATCF data that only has 8 columns?
If
> so, perhaps that we should change the minimum required from 18 to 8.
>
> John
>
> On Wed, Jun 24, 2015 at 4:42 AM, Andrew Penny - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > Wed Jun 24 04:42:49 2015: Request 72431 was acted upon.
> > Transaction: Ticket created by andrew.penny at noaa.gov
> >        Queue: met_help
> >      Subject: suggestion for tc_pairs
> >        Owner: Nobody
> >   Requestors: andrew.penny at noaa.gov
> >       Status: new
> >  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
> >
> >
> > Hi there,
> >
> > Just a quick suggestion for tc_pairs. When using tc_pairs, if
there are
> > less than 18 fields in the ATCF file, tc_pairs will issue the
following
> > warning:
> >
> > WARNING:
> > WARNING: bool operator>>(istream & in, ATCFLine & t) -> found
fewer than
> > the expected number of elements (8<18) in ATCF line:
> > WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
> > WARNING:
> >
> > ​and then stop reading the file. ​This is somewhat of a pain in
that the
> > majority of the ATCF files I need to verify do not have a full 18
> > columns, and I need to pre-process them all to ensure tc_pairs
does not
> > complain. Can a flag option be included in the namelist or
> > configure file to automatically fill out a ATCF file with null
values if
> > the number of columns is below the needed amount? I don't know
> > how much work this would entail, but it would be a great option to
have.
> > tc_pairs will currently exit if it encounters a non-standard line,
> > even if it doesn't correspond to one of the models you're trying
to
> verify.
> >
> > Aside from this rather small issue, I really appreciate all of the
> > functionality available with tc_pairs and tc_stat. Thanks for all
your
> hard
> > work!
> >
> > Take care,
> > Andy
> >
> > --
> > Andrew Penny
> > Hurricane Model Diagnostician
> > SRG / National Hurricane Center
> > 11691 SW 17th Street
> > Miami, FL 33165-2149
> > phone: 305.229.4457
> > email: andrew.penny at noaa.gov
> >
> >
>
>


--
Andrew Penny
Hurricane Model Diagnostician
SRG / National Hurricane Center
11691 SW 17th Street
Miami, FL 33165-2149
phone: 305.229.4457
email: andrew.penny at noaa.gov

------------------------------------------------
Subject: suggestion for tc_pairs
From: John Halley Gotway
Time: Thu Jun 25 09:45:17 2015

Andrew,

Thanks for letting me know.  I defined a new development task to
refine
that logic for the next release of MET.

Thanks,
John

On Wed, Jun 24, 2015 at 10:35 AM, Andrew Penny - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
>
> Hi John,
>
> I've recently been doing a lot of verification for retrospective
runs (to
> evaluate proposed model upgrades). To be honest, I'm not sure why
some ATCF
> files are more filled out than others. Perhaps the other fields are
left
> out to save time since just the track and intensity are used to
verify the
> retrospective runs. In fact, sometimes there isn't a column for
intensity,
> as some models are not used for intensity guidance.
>
> Hope that helps. Let me know if you need more information.
>
> Thanks and take care,
> Andy
>
> On Wed, Jun 24, 2015 at 11:30 AM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Andy,
> >
> > Sure, that'd be simple enough.  I'll add it to our list of
development
> > tasks.  Can you tell me though, what is the source of this ATCF
data that
> > you're using?  Is it real hurricane data or is it some novel use
of the
> > ATCF file format?  Knowing that would help put your request in
context.
> Is
> > someone out there generating real ATCF data that only has 8
columns?  If
> > so, perhaps that we should change the minimum required from 18 to
8.
> >
> > John
> >
> > On Wed, Jun 24, 2015 at 4:42 AM, Andrew Penny - NOAA Affiliate via
RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > Wed Jun 24 04:42:49 2015: Request 72431 was acted upon.
> > > Transaction: Ticket created by andrew.penny at noaa.gov
> > >        Queue: met_help
> > >      Subject: suggestion for tc_pairs
> > >        Owner: Nobody
> > >   Requestors: andrew.penny at noaa.gov
> > >       Status: new
> > >  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431
> >
> > >
> > >
> > > Hi there,
> > >
> > > Just a quick suggestion for tc_pairs. When using tc_pairs, if
there are
> > > less than 18 fields in the ATCF file, tc_pairs will issue the
following
> > > warning:
> > >
> > > WARNING:
> > > WARNING: bool operator>>(istream & in, ATCFLine & t) -> found
fewer
> than
> > > the expected number of elements (8<18) in ATCF line:
> > > WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
> > > WARNING:
> > >
> > > ​and then stop reading the file. ​This is somewhat of a pain in
that
> the
> > > majority of the ATCF files I need to verify do not have a full
18
> > > columns, and I need to pre-process them all to ensure tc_pairs
does not
> > > complain. Can a flag option be included in the namelist or
> > > configure file to automatically fill out a ATCF file with null
values
> if
> > > the number of columns is below the needed amount? I don't know
> > > how much work this would entail, but it would be a great option
to
> have.
> > > tc_pairs will currently exit if it encounters a non-standard
line,
> > > even if it doesn't correspond to one of the models you're trying
to
> > verify.
> > >
> > > Aside from this rather small issue, I really appreciate all of
the
> > > functionality available with tc_pairs and tc_stat. Thanks for
all your
> > hard
> > > work!
> > >
> > > Take care,
> > > Andy
> > >
> > > --
> > > Andrew Penny
> > > Hurricane Model Diagnostician
> > > SRG / National Hurricane Center
> > > 11691 SW 17th Street
> > > Miami, FL 33165-2149
> > > phone: 305.229.4457
> > > email: andrew.penny at noaa.gov
> > >
> > >
> >
> >
>
>
> --
> Andrew Penny
> Hurricane Model Diagnostician
> SRG / National Hurricane Center
> 11691 SW 17th Street
> Miami, FL 33165-2149
> phone: 305.229.4457
> email: andrew.penny at noaa.gov
>
>

------------------------------------------------
Subject: suggestion for tc_pairs
From: Andrew Penny - NOAA Affiliate
Time: Fri Jul 17 06:38:08 2015

Hi John,

I have a few more questions regarding MET-TC. I'm not currently
getting any
output (just N/A or 0) in the "independence time" and "frequency of
superior performance" columns: TS_INT, TS_IND, FSP_TOTAL, FSP_BEST,
FSP_TIES, FSP. Is there something I need to set in the configure file
to
have these computed?

Also, are there file naming requirements for running tc_pairs? It
seems
that unless my atcf file ends in ".dat", tc_pairs won't recognize the
file
properly.

Thanks,
Andy

On Thu, Jun 25, 2015 at 11:45 AM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:

> Andrew,
>
> Thanks for letting me know.  I defined a new development task to
refine
> that logic for the next release of MET.
>
> Thanks,
> John
>
> On Wed, Jun 24, 2015 at 10:35 AM, Andrew Penny - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
> >
> > Hi John,
> >
> > I've recently been doing a lot of verification for retrospective
runs (to
> > evaluate proposed model upgrades). To be honest, I'm not sure why
some
> ATCF
> > files are more filled out than others. Perhaps the other fields
are left
> > out to save time since just the track and intensity are used to
verify
> the
> > retrospective runs. In fact, sometimes there isn't a column for
> intensity,
> > as some models are not used for intensity guidance.
> >
> > Hope that helps. Let me know if you need more information.
> >
> > Thanks and take care,
> > Andy
> >
> > On Wed, Jun 24, 2015 at 11:30 AM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Andy,
> > >
> > > Sure, that'd be simple enough.  I'll add it to our list of
development
> > > tasks.  Can you tell me though, what is the source of this ATCF
data
> that
> > > you're using?  Is it real hurricane data or is it some novel use
of the
> > > ATCF file format?  Knowing that would help put your request in
context.
> > Is
> > > someone out there generating real ATCF data that only has 8
columns?
> If
> > > so, perhaps that we should change the minimum required from 18
to 8.
> > >
> > > John
> > >
> > > On Wed, Jun 24, 2015 at 4:42 AM, Andrew Penny - NOAA Affiliate
via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > Wed Jun 24 04:42:49 2015: Request 72431 was acted upon.
> > > > Transaction: Ticket created by andrew.penny at noaa.gov
> > > >        Queue: met_help
> > > >      Subject: suggestion for tc_pairs
> > > >        Owner: Nobody
> > > >   Requestors: andrew.penny at noaa.gov
> > > >       Status: new
> > > >  Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431
> > >
> > > >
> > > >
> > > > Hi there,
> > > >
> > > > Just a quick suggestion for tc_pairs. When using tc_pairs, if
there
> are
> > > > less than 18 fields in the ATCF file, tc_pairs will issue the
> following
> > > > warning:
> > > >
> > > > WARNING:
> > > > WARNING: bool operator>>(istream & in, ATCFLine & t) -> found
fewer
> > than
> > > > the expected number of elements (8<18) in ATCF line:
> > > > WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
> > > > WARNING:
> > > >
> > > > ​and then stop reading the file. ​This is somewhat of a pain
in that
> > the
> > > > majority of the ATCF files I need to verify do not have a full
18
> > > > columns, and I need to pre-process them all to ensure tc_pairs
does
> not
> > > > complain. Can a flag option be included in the namelist or
> > > > configure file to automatically fill out a ATCF file with null
values
> > if
> > > > the number of columns is below the needed amount? I don't know
> > > > how much work this would entail, but it would be a great
option to
> > have.
> > > > tc_pairs will currently exit if it encounters a non-standard
line,
> > > > even if it doesn't correspond to one of the models you're
trying to
> > > verify.
> > > >
> > > > Aside from this rather small issue, I really appreciate all of
the
> > > > functionality available with tc_pairs and tc_stat. Thanks for
all
> your
> > > hard
> > > > work!
> > > >
> > > > Take care,
> > > > Andy
> > > >
> > > > --
> > > > Andrew Penny
> > > > Hurricane Model Diagnostician
> > > > SRG / National Hurricane Center
> > > > 11691 SW 17th Street
> > > > Miami, FL 33165-2149
> > > > phone: 305.229.4457
> > > > email: andrew.penny at noaa.gov
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Andrew Penny
> > Hurricane Model Diagnostician
> > SRG / National Hurricane Center
> > 11691 SW 17th Street
> > Miami, FL 33165-2149
> > phone: 305.229.4457
> > email: andrew.penny at noaa.gov
> >
> >
>
>


--
Andrew Penny
Hurricane Model Diagnostician
SRG / National Hurricane Center
11691 SW 17th Street
Miami, FL 33165-2149
phone: 305.229.4457
email: andrew.penny at noaa.gov

------------------------------------------------
Subject: suggestion for tc_pairs
From: John Halley Gotway
Time: Fri Jul 17 15:06:09 2015

Andy,

Yes, that's correct.  In met-5.0, tc_pairs only searches for files
ending
in .dat, as defined for us by NHC.  However, I can see a way we could
nicely generalize this.  If the user passes us a directory in "-adeck"
or
"-bdeck" we could continue looking for the .dat extension.  But if
they
pass us explicit file names, we could read them directly regardless of
the
the file's extension.  Is that the sort of behavior you're looking
for?

We actually made a very similar change to the MET stat_analysis tool
for
the version 5.1 release.  If the user passes in a directory, we
recursively
search for files ending in ".stat".  But if they pass in a file name,
we
just read it directly.

Computing the frequency of superior performance is pretty finicky.  It
can
only be computed when event equalization is turned on.  I've attached
a
sample output file from tc_pairs.  Try using it to run the following
tc_stat jobs:

(1) Notice that this produces no output since there are no cases
common to
all the models in my dataset:
   tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
-event_equal true

(2) Limiting the analysis to 2 models (AHWI and BCLP) enables FSP to
be
computed:
   tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
-event_equal true -amodel AHWI,BCLP

(3) Try adding "-by LEAD" to break it out by lead time and there are
no
results again:
   tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL,LEAD -event_equal true -amodel AHWI,BCLP

(4) Instead, fix a specific lead time and FSP is valid again:
   tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
-lead 48 -event_equal true -amodel AHWI,BCLP

But you don't get time series info unless you specify a single
initialization time:
   tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
-event_equal true -amodel AHWI,BCLP -init_inc 20101107_180000

Clear as mud?

John





On Fri, Jul 17, 2015 at 6:38 AM, Andrew Penny - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
>
> Hi John,
>
> I have a few more questions regarding MET-TC. I'm not currently
getting any
> output (just N/A or 0) in the "independence time" and "frequency of
> superior performance" columns: TS_INT, TS_IND, FSP_TOTAL, FSP_BEST,
> FSP_TIES, FSP. Is there something I need to set in the configure
file to
> have these computed?
>
> Also, are there file naming requirements for running tc_pairs? It
seems
> that unless my atcf file ends in ".dat", tc_pairs won't recognize
the file
> properly.
>
> Thanks,
> Andy
>
> On Thu, Jun 25, 2015 at 11:45 AM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Andrew,
> >
> > Thanks for letting me know.  I defined a new development task to
refine
> > that logic for the next release of MET.
> >
> > Thanks,
> > John
> >
> > On Wed, Jun 24, 2015 at 10:35 AM, Andrew Penny - NOAA Affiliate
via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
> > >
> > > Hi John,
> > >
> > > I've recently been doing a lot of verification for retrospective
runs
> (to
> > > evaluate proposed model upgrades). To be honest, I'm not sure
why some
> > ATCF
> > > files are more filled out than others. Perhaps the other fields
are
> left
> > > out to save time since just the track and intensity are used to
verify
> > the
> > > retrospective runs. In fact, sometimes there isn't a column for
> > intensity,
> > > as some models are not used for intensity guidance.
> > >
> > > Hope that helps. Let me know if you need more information.
> > >
> > > Thanks and take care,
> > > Andy
> > >
> > > On Wed, Jun 24, 2015 at 11:30 AM, John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Andy,
> > > >
> > > > Sure, that'd be simple enough.  I'll add it to our list of
> development
> > > > tasks.  Can you tell me though, what is the source of this
ATCF data
> > that
> > > > you're using?  Is it real hurricane data or is it some novel
use of
> the
> > > > ATCF file format?  Knowing that would help put your request in
> context.
> > > Is
> > > > someone out there generating real ATCF data that only has 8
columns?
> > If
> > > > so, perhaps that we should change the minimum required from 18
to 8.
> > > >
> > > > John
> > > >
> > > > On Wed, Jun 24, 2015 at 4:42 AM, Andrew Penny - NOAA Affiliate
via
> RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > Wed Jun 24 04:42:49 2015: Request 72431 was acted upon.
> > > > > Transaction: Ticket created by andrew.penny at noaa.gov
> > > > >        Queue: met_help
> > > > >      Subject: suggestion for tc_pairs
> > > > >        Owner: Nobody
> > > > >   Requestors: andrew.penny at noaa.gov
> > > > >       Status: new
> > > > >  Ticket <URL:
> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431
> > > >
> > > > >
> > > > >
> > > > > Hi there,
> > > > >
> > > > > Just a quick suggestion for tc_pairs. When using tc_pairs,
if there
> > are
> > > > > less than 18 fields in the ATCF file, tc_pairs will issue
the
> > following
> > > > > warning:
> > > > >
> > > > > WARNING:
> > > > > WARNING: bool operator>>(istream & in, ATCFLine & t) ->
found fewer
> > > than
> > > > > the expected number of elements (8<18) in ATCF line:
> > > > > WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
> > > > > WARNING:
> > > > >
> > > > > ​and then stop reading the file. ​This is somewhat of a pain
in
> that
> > > the
> > > > > majority of the ATCF files I need to verify do not have a
full 18
> > > > > columns, and I need to pre-process them all to ensure
tc_pairs does
> > not
> > > > > complain. Can a flag option be included in the namelist or
> > > > > configure file to automatically fill out a ATCF file with
null
> values
> > > if
> > > > > the number of columns is below the needed amount? I don't
know
> > > > > how much work this would entail, but it would be a great
option to
> > > have.
> > > > > tc_pairs will currently exit if it encounters a non-standard
line,
> > > > > even if it doesn't correspond to one of the models you're
trying to
> > > > verify.
> > > > >
> > > > > Aside from this rather small issue, I really appreciate all
of the
> > > > > functionality available with tc_pairs and tc_stat. Thanks
for all
> > your
> > > > hard
> > > > > work!
> > > > >
> > > > > Take care,
> > > > > Andy
> > > > >
> > > > > --
> > > > > Andrew Penny
> > > > > Hurricane Model Diagnostician
> > > > > SRG / National Hurricane Center
> > > > > 11691 SW 17th Street
> > > > > Miami, FL 33165-2149
> > > > > phone: 305.229.4457
> > > > > email: andrew.penny at noaa.gov
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Andrew Penny
> > > Hurricane Model Diagnostician
> > > SRG / National Hurricane Center
> > > 11691 SW 17th Street
> > > Miami, FL 33165-2149
> > > phone: 305.229.4457
> > > email: andrew.penny at noaa.gov
> > >
> > >
> >
> >
>
>
> --
> Andrew Penny
> Hurricane Model Diagnostician
> SRG / National Hurricane Center
> 11691 SW 17th Street
> Miami, FL 33165-2149
> phone: 305.229.4457
> email: andrew.penny at noaa.gov
>
>

------------------------------------------------
Subject: suggestion for tc_pairs
From: Andrew Penny - NOAA Affiliate
Time: Tue Jul 21 10:55:31 2015

Hi John,

Yes, that's the type of behavior I was thinking of regarding the input
file
names for tc_pairs. It would be helpful as there are many times that
the
input files don't follow the strict ".dat" naming conventions when
we're
evaluating/verifying non-operational model output.

Thanks for the explanation regarding the frequency of superior
performance.
I will do some experimenting with the file you sent, and may have some
more
follow-on questions.

Thanks for your help.

Take care,
Andy

On Fri, Jul 17, 2015 at 5:06 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:

> Andy,
>
> Yes, that's correct.  In met-5.0, tc_pairs only searches for files
ending
> in .dat, as defined for us by NHC.  However, I can see a way we
could
> nicely generalize this.  If the user passes us a directory in "-
adeck" or
> "-bdeck" we could continue looking for the .dat extension.  But if
they
> pass us explicit file names, we could read them directly regardless
of the
> the file's extension.  Is that the sort of behavior you're looking
for?
>
> We actually made a very similar change to the MET stat_analysis tool
for
> the version 5.1 release.  If the user passes in a directory, we
recursively
> search for files ending in ".stat".  But if they pass in a file
name, we
> just read it directly.
>
> Computing the frequency of superior performance is pretty finicky.
It can
> only be computed when event equalization is turned on.  I've
attached a
> sample output file from tc_pairs.  Try using it to run the following
> tc_stat jobs:
>
> (1) Notice that this produces no output since there are no cases
common to
> all the models in my dataset:
>    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> -event_equal true
>
> (2) Limiting the analysis to 2 models (AHWI and BCLP) enables FSP to
be
> computed:
>    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> -event_equal true -amodel AHWI,BCLP
>
> (3) Try adding "-by LEAD" to break it out by lead time and there are
no
> results again:
>    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
> AMODEL,LEAD -event_equal true -amodel AHWI,BCLP
>
> (4) Instead, fix a specific lead time and FSP is valid again:
>    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> -lead 48 -event_equal true -amodel AHWI,BCLP
>
> But you don't get time series info unless you specify a single
> initialization time:
>    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> -event_equal true -amodel AHWI,BCLP -init_inc 20101107_180000
>
> Clear as mud?
>
> John
>
>
>
>
>
> On Fri, Jul 17, 2015 at 6:38 AM, Andrew Penny - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
> >
> > Hi John,
> >
> > I have a few more questions regarding MET-TC. I'm not currently
getting
> any
> > output (just N/A or 0) in the "independence time" and "frequency
of
> > superior performance" columns: TS_INT, TS_IND, FSP_TOTAL,
FSP_BEST,
> > FSP_TIES, FSP. Is there something I need to set in the configure
file to
> > have these computed?
> >
> > Also, are there file naming requirements for running tc_pairs? It
seems
> > that unless my atcf file ends in ".dat", tc_pairs won't recognize
the
> file
> > properly.
> >
> > Thanks,
> > Andy
> >
> > On Thu, Jun 25, 2015 at 11:45 AM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Andrew,
> > >
> > > Thanks for letting me know.  I defined a new development task to
refine
> > > that logic for the next release of MET.
> > >
> > > Thanks,
> > > John
> > >
> > > On Wed, Jun 24, 2015 at 10:35 AM, Andrew Penny - NOAA Affiliate
via RT
> <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431
>
> > > >
> > > > Hi John,
> > > >
> > > > I've recently been doing a lot of verification for
retrospective runs
> > (to
> > > > evaluate proposed model upgrades). To be honest, I'm not sure
why
> some
> > > ATCF
> > > > files are more filled out than others. Perhaps the other
fields are
> > left
> > > > out to save time since just the track and intensity are used
to
> verify
> > > the
> > > > retrospective runs. In fact, sometimes there isn't a column
for
> > > intensity,
> > > > as some models are not used for intensity guidance.
> > > >
> > > > Hope that helps. Let me know if you need more information.
> > > >
> > > > Thanks and take care,
> > > > Andy
> > > >
> > > > On Wed, Jun 24, 2015 at 11:30 AM, John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > > Andy,
> > > > >
> > > > > Sure, that'd be simple enough.  I'll add it to our list of
> > development
> > > > > tasks.  Can you tell me though, what is the source of this
ATCF
> data
> > > that
> > > > > you're using?  Is it real hurricane data or is it some novel
use of
> > the
> > > > > ATCF file format?  Knowing that would help put your request
in
> > context.
> > > > Is
> > > > > someone out there generating real ATCF data that only has 8
> columns?
> > > If
> > > > > so, perhaps that we should change the minimum required from
18 to
> 8.
> > > > >
> > > > > John
> > > > >
> > > > > On Wed, Jun 24, 2015 at 4:42 AM, Andrew Penny - NOAA
Affiliate via
> > RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > >
> > > > > > Wed Jun 24 04:42:49 2015: Request 72431 was acted upon.
> > > > > > Transaction: Ticket created by andrew.penny at noaa.gov
> > > > > >        Queue: met_help
> > > > > >      Subject: suggestion for tc_pairs
> > > > > >        Owner: Nobody
> > > > > >   Requestors: andrew.penny at noaa.gov
> > > > > >       Status: new
> > > > > >  Ticket <URL:
> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431
> > > > >
> > > > > >
> > > > > >
> > > > > > Hi there,
> > > > > >
> > > > > > Just a quick suggestion for tc_pairs. When using tc_pairs,
if
> there
> > > are
> > > > > > less than 18 fields in the ATCF file, tc_pairs will issue
the
> > > following
> > > > > > warning:
> > > > > >
> > > > > > WARNING:
> > > > > > WARNING: bool operator>>(istream & in, ATCFLine & t) ->
found
> fewer
> > > > than
> > > > > > the expected number of elements (8<18) in ATCF line:
> > > > > > WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
> > > > > > WARNING:
> > > > > >
> > > > > > ​and then stop reading the file. ​This is somewhat of a
pain in
> > that
> > > > the
> > > > > > majority of the ATCF files I need to verify do not have a
full 18
> > > > > > columns, and I need to pre-process them all to ensure
tc_pairs
> does
> > > not
> > > > > > complain. Can a flag option be included in the namelist or
> > > > > > configure file to automatically fill out a ATCF file with
null
> > values
> > > > if
> > > > > > the number of columns is below the needed amount? I don't
know
> > > > > > how much work this would entail, but it would be a great
option
> to
> > > > have.
> > > > > > tc_pairs will currently exit if it encounters a non-
standard
> line,
> > > > > > even if it doesn't correspond to one of the models you're
trying
> to
> > > > > verify.
> > > > > >
> > > > > > Aside from this rather small issue, I really appreciate
all of
> the
> > > > > > functionality available with tc_pairs and tc_stat. Thanks
for all
> > > your
> > > > > hard
> > > > > > work!
> > > > > >
> > > > > > Take care,
> > > > > > Andy
> > > > > >
> > > > > > --
> > > > > > Andrew Penny
> > > > > > Hurricane Model Diagnostician
> > > > > > SRG / National Hurricane Center
> > > > > > 11691 SW 17th Street
> > > > > > Miami, FL 33165-2149
> > > > > > phone: 305.229.4457
> > > > > > email: andrew.penny at noaa.gov
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Andrew Penny
> > > > Hurricane Model Diagnostician
> > > > SRG / National Hurricane Center
> > > > 11691 SW 17th Street
> > > > Miami, FL 33165-2149
> > > > phone: 305.229.4457
> > > > email: andrew.penny at noaa.gov
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Andrew Penny
> > Hurricane Model Diagnostician
> > SRG / National Hurricane Center
> > 11691 SW 17th Street
> > Miami, FL 33165-2149
> > phone: 305.229.4457
> > email: andrew.penny at noaa.gov
> >
> >
>
>


--
Andrew Penny
Hurricane Model Diagnostician
SRG / National Hurricane Center
11691 SW 17th Street
Miami, FL 33165-2149
phone: 305.229.4457
email: andrew.penny at noaa.gov

------------------------------------------------
Subject: suggestion for tc_pairs
From: John Halley Gotway
Time: Tue Jul 21 13:24:11 2015

Andy,

Thanks for getting back to me on this.  I created a development ticket
describing the change in functionality for tc_pairs that we discussed.
If
possible, we'll get it into the next release, but I can't make any
promises
on that.

I'll go ahead and resolve this ticket now, but feel free to write with
any
more issues or questions that come up.

Thanks,
John Halley Gotway

On Tue, Jul 21, 2015 at 10:55 AM, Andrew Penny - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
>
> Hi John,
>
> Yes, that's the type of behavior I was thinking of regarding the
input file
> names for tc_pairs. It would be helpful as there are many times that
the
> input files don't follow the strict ".dat" naming conventions when
we're
> evaluating/verifying non-operational model output.
>
> Thanks for the explanation regarding the frequency of superior
performance.
> I will do some experimenting with the file you sent, and may have
some more
> follow-on questions.
>
> Thanks for your help.
>
> Take care,
> Andy
>
> On Fri, Jul 17, 2015 at 5:06 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Andy,
> >
> > Yes, that's correct.  In met-5.0, tc_pairs only searches for files
ending
> > in .dat, as defined for us by NHC.  However, I can see a way we
could
> > nicely generalize this.  If the user passes us a directory in "-
adeck" or
> > "-bdeck" we could continue looking for the .dat extension.  But if
they
> > pass us explicit file names, we could read them directly
regardless of
> the
> > the file's extension.  Is that the sort of behavior you're looking
for?
> >
> > We actually made a very similar change to the MET stat_analysis
tool for
> > the version 5.1 release.  If the user passes in a directory, we
> recursively
> > search for files ending in ".stat".  But if they pass in a file
name, we
> > just read it directly.
> >
> > Computing the frequency of superior performance is pretty finicky.
It
> can
> > only be computed when event equalization is turned on.  I've
attached a
> > sample output file from tc_pairs.  Try using it to run the
following
> > tc_stat jobs:
> >
> > (1) Notice that this produces no output since there are no cases
common
> to
> > all the models in my dataset:
> >    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> > -event_equal true
> >
> > (2) Limiting the analysis to 2 models (AHWI and BCLP) enables FSP
to be
> > computed:
> >    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> > -event_equal true -amodel AHWI,BCLP
> >
> > (3) Try adding "-by LEAD" to break it out by lead time and there
are no
> > results again:
> >    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
> > AMODEL,LEAD -event_equal true -amodel AHWI,BCLP
> >
> > (4) Instead, fix a specific lead time and FSP is valid again:
> >    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> > -lead 48 -event_equal true -amodel AHWI,BCLP
> >
> > But you don't get time series info unless you specify a single
> > initialization time:
> >    tc_stat -lookin alal2010.tcst -job summary -column TK_ERR -by
AMODEL
> > -event_equal true -amodel AHWI,BCLP -init_inc 20101107_180000
> >
> > Clear as mud?
> >
> > John
> >
> >
> >
> >
> >
> > On Fri, Jul 17, 2015 at 6:38 AM, Andrew Penny - NOAA Affiliate via
RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
> > >
> > > Hi John,
> > >
> > > I have a few more questions regarding MET-TC. I'm not currently
getting
> > any
> > > output (just N/A or 0) in the "independence time" and "frequency
of
> > > superior performance" columns: TS_INT, TS_IND, FSP_TOTAL,
FSP_BEST,
> > > FSP_TIES, FSP. Is there something I need to set in the configure
file
> to
> > > have these computed?
> > >
> > > Also, are there file naming requirements for running tc_pairs?
It seems
> > > that unless my atcf file ends in ".dat", tc_pairs won't
recognize the
> > file
> > > properly.
> > >
> > > Thanks,
> > > Andy
> > >
> > > On Thu, Jun 25, 2015 at 11:45 AM, John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Andrew,
> > > >
> > > > Thanks for letting me know.  I defined a new development task
to
> refine
> > > > that logic for the next release of MET.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Wed, Jun 24, 2015 at 10:35 AM, Andrew Penny - NOAA
Affiliate via
> RT
> > <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431 >
> > > > >
> > > > > Hi John,
> > > > >
> > > > > I've recently been doing a lot of verification for
retrospective
> runs
> > > (to
> > > > > evaluate proposed model upgrades). To be honest, I'm not
sure why
> > some
> > > > ATCF
> > > > > files are more filled out than others. Perhaps the other
fields are
> > > left
> > > > > out to save time since just the track and intensity are used
to
> > verify
> > > > the
> > > > > retrospective runs. In fact, sometimes there isn't a column
for
> > > > intensity,
> > > > > as some models are not used for intensity guidance.
> > > > >
> > > > > Hope that helps. Let me know if you need more information.
> > > > >
> > > > > Thanks and take care,
> > > > > Andy
> > > > >
> > > > > On Wed, Jun 24, 2015 at 11:30 AM, John Halley Gotway via RT
<
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > > Andy,
> > > > > >
> > > > > > Sure, that'd be simple enough.  I'll add it to our list of
> > > development
> > > > > > tasks.  Can you tell me though, what is the source of this
ATCF
> > data
> > > > that
> > > > > > you're using?  Is it real hurricane data or is it some
novel use
> of
> > > the
> > > > > > ATCF file format?  Knowing that would help put your
request in
> > > context.
> > > > > Is
> > > > > > someone out there generating real ATCF data that only has
8
> > columns?
> > > > If
> > > > > > so, perhaps that we should change the minimum required
from 18 to
> > 8.
> > > > > >
> > > > > > John
> > > > > >
> > > > > > On Wed, Jun 24, 2015 at 4:42 AM, Andrew Penny - NOAA
Affiliate
> via
> > > RT <
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > > >
> > > > > > > Wed Jun 24 04:42:49 2015: Request 72431 was acted upon.
> > > > > > > Transaction: Ticket created by andrew.penny at noaa.gov
> > > > > > >        Queue: met_help
> > > > > > >      Subject: suggestion for tc_pairs
> > > > > > >        Owner: Nobody
> > > > > > >   Requestors: andrew.penny at noaa.gov
> > > > > > >       Status: new
> > > > > > >  Ticket <URL:
> > > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=72431
> > > > > >
> > > > > > >
> > > > > > >
> > > > > > > Hi there,
> > > > > > >
> > > > > > > Just a quick suggestion for tc_pairs. When using
tc_pairs, if
> > there
> > > > are
> > > > > > > less than 18 fields in the ATCF file, tc_pairs will
issue the
> > > > following
> > > > > > > warning:
> > > > > > >
> > > > > > > WARNING:
> > > > > > > WARNING: bool operator>>(istream & in, ATCFLine & t) ->
found
> > fewer
> > > > > than
> > > > > > > the expected number of elements (8<18) in ATCF line:
> > > > > > > WARNING: AL, 01, 2011062900, 03, T254, 000, 211N,  937W
> > > > > > > WARNING:
> > > > > > >
> > > > > > > ​and then stop reading the file. ​This is somewhat of a
pain in
> > > that
> > > > > the
> > > > > > > majority of the ATCF files I need to verify do not have
a full
> 18
> > > > > > > columns, and I need to pre-process them all to ensure
tc_pairs
> > does
> > > > not
> > > > > > > complain. Can a flag option be included in the namelist
or
> > > > > > > configure file to automatically fill out a ATCF file
with null
> > > values
> > > > > if
> > > > > > > the number of columns is below the needed amount? I
don't know
> > > > > > > how much work this would entail, but it would be a great
option
> > to
> > > > > have.
> > > > > > > tc_pairs will currently exit if it encounters a non-
standard
> > line,
> > > > > > > even if it doesn't correspond to one of the models
you're
> trying
> > to
> > > > > > verify.
> > > > > > >
> > > > > > > Aside from this rather small issue, I really appreciate
all of
> > the
> > > > > > > functionality available with tc_pairs and tc_stat.
Thanks for
> all
> > > > your
> > > > > > hard
> > > > > > > work!
> > > > > > >
> > > > > > > Take care,
> > > > > > > Andy
> > > > > > >
> > > > > > > --
> > > > > > > Andrew Penny
> > > > > > > Hurricane Model Diagnostician
> > > > > > > SRG / National Hurricane Center
> > > > > > > 11691 SW 17th Street
> > > > > > > Miami, FL 33165-2149
> > > > > > > phone: 305.229.4457
> > > > > > > email: andrew.penny at noaa.gov
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Andrew Penny
> > > > > Hurricane Model Diagnostician
> > > > > SRG / National Hurricane Center
> > > > > 11691 SW 17th Street
> > > > > Miami, FL 33165-2149
> > > > > phone: 305.229.4457
> > > > > email: andrew.penny at noaa.gov
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Andrew Penny
> > > Hurricane Model Diagnostician
> > > SRG / National Hurricane Center
> > > 11691 SW 17th Street
> > > Miami, FL 33165-2149
> > > phone: 305.229.4457
> > > email: andrew.penny at noaa.gov
> > >
> > >
> >
> >
>
>
> --
> Andrew Penny
> Hurricane Model Diagnostician
> SRG / National Hurricane Center
> 11691 SW 17th Street
> Miami, FL 33165-2149
> phone: 305.229.4457
> email: andrew.penny at noaa.gov
>
>

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


More information about the Met_help mailing list