[Met_help] [rt.rap.ucar.edu #68989] History for rapid_intens filtering questions with tc_stat in MET-TC
John Halley Gotway via RT
met_help at ucar.edu
Fri Oct 3 12:23:50 MDT 2014
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi,
I was wondering if I could get clarification on how applying the
"rapid_intens" flag in tc_stat functions.
I was under the impression that it would find periods of rapid
intensification in the b-deck, and then only select forecast times that
applied to the previous 24 hours (or perhaps earlier).
EX: Hurricane John undergoes rapid intensification (RI) from 00Z on the
21st to 00Z on the 22nd.
in the b-deck, WIND(00Z 22nd) - WIND(00Z 21st) >= 30 kts, so RI achieved.
Therefore, tc_stat should keep all forecasts from 00Z on the 21st (and
perhaps 00Z on the 20th and 00Z on the 19th, etc.) depending on whether you
want the forecast from where RI began or all forecasts which include lead
times where the RI occured.
However, if I apply the rapid_intens criteria (on Yellowstone in
/glade/p/work/zarzycki/test_met/) when just testing the one of the provided
R scripts and some random models...
Rscript /glade/u/home/zarzycki/METv4.1/scripts/Rscripts/plot_tcmpr.R
-lookin /glade/u/home/zarzycki/work/test_met/out_tcmpr.tcst -filter
"-amodel EGRR,GFDL,CMZ,GFSO,HWRF -rapid_inten true -rapid_inten_thresh
ge30" -save_data "./save_the_data.out" -dep "TK_ERR,AMAX_WIND-BMAX_WIND"
-series AMODEL EGRR,GFDL,CMZ,GFSO,HWRF -plot MEAN,BOXPLOT -outdir
/glade/u/home/zarzycki/work/test_met/ -lead 12,24,48,72,96,120,144,168,192
-no_ee
I get A) a tremendously large number of matches for our model and not other
models at 24 hour lead, which shouldn't be the case if we are only using
b-deck (these numbers should be similar, since, generally, the cases all
have the same init times).
B) the matches for other models (ex: GFSO, etc.) include storms like
Florence (2012) which is an example of a non-RI case (it was a weak
tropical storm that dissipated within a few days).
Am I doing something incorrect when setting this flag? The number of storms
that DID undergo RI during our sampling period was relatively small, so it
may be a moot point, but this behavior seems odd.
Cheers,
-Colin
--
Colin M. Zarzycki, ASP postdoctoral fellow
Atmospheric Modeling and Predictability
National Center for Atmospheric Research
http://www.colinzarzycki.com
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: rapid_intens filtering questions with tc_stat in MET-TC
From: Colin Zarzycki
Time: Tue Sep 16 12:06:20 2014
It appears many of my issues are mitigated by moving to MET 5.0. The
results are at least somewhat consistent between the models
using -rapid_inten_track BDECK (which I believe is the behavior I
want).
Of note, the 5.0 documentation does not document the command line
implementation of the new RI settings (I had to peek at the source
code to
note that each flag (-rapid_inten_track, rapid_intens_thresh, etc.)
needed
to be specified as opposed to one rapid_inten setting).
It still remains unclear to me how RI is defined, but I will play
around
with it a bit first.
-Colin
On Tue, Sep 16, 2014 at 10:28 AM, met_help at ucar.edu via RT <
met_help at ucar.edu> wrote:
> Greetings,
>
> This message has been automatically generated in response to the
> creation of a trouble ticket regarding:
> "rapid_intens filtering questions with tc_stat in MET-TC",
> a summary of which appears below.
>
> There is no need to reply to this message right now. Your ticket
has been
> assigned an ID of [rt.rap.ucar.edu #68989].
>
> Please include the string:
>
> [rt.rap.ucar.edu #68989]
>
> in the subject line of all future correspondence about this issue.
To do
> so,
> you may reply to this message.
>
> Thank you,
> met_help at ucar.edu
>
>
-------------------------------------------------------------------------
> Hi,
>
> I was wondering if I could get clarification on how applying the
> "rapid_intens" flag in tc_stat functions.
>
> I was under the impression that it would find periods of rapid
> intensification in the b-deck, and then only select forecast times
that
> applied to the previous 24 hours (or perhaps earlier).
>
> EX: Hurricane John undergoes rapid intensification (RI) from 00Z on
the
> 21st to 00Z on the 22nd.
> in the b-deck, WIND(00Z 22nd) - WIND(00Z 21st) >= 30 kts, so RI
achieved.
> Therefore, tc_stat should keep all forecasts from 00Z on the 21st
(and
> perhaps 00Z on the 20th and 00Z on the 19th, etc.) depending on
whether you
> want the forecast from where RI began or all forecasts which include
lead
> times where the RI occured.
>
> However, if I apply the rapid_intens criteria (on Yellowstone in
> /glade/p/work/zarzycki/test_met/) when just testing the one of the
provided
> R scripts and some random models...
>
> Rscript /glade/u/home/zarzycki/METv4.1/scripts/Rscripts/plot_tcmpr.R
> -lookin /glade/u/home/zarzycki/work/test_met/out_tcmpr.tcst -filter
> "-amodel EGRR,GFDL,CMZ,GFSO,HWRF -rapid_inten true
-rapid_inten_thresh
> ge30" -save_data "./save_the_data.out" -dep "TK_ERR,AMAX_WIND-
BMAX_WIND"
> -series AMODEL EGRR,GFDL,CMZ,GFSO,HWRF -plot MEAN,BOXPLOT -outdir
> /glade/u/home/zarzycki/work/test_met/ -lead
12,24,48,72,96,120,144,168,192
> -no_ee
>
> I get A) a tremendously large number of matches for our model and
not other
> models at 24 hour lead, which shouldn't be the case if we are only
using
> b-deck (these numbers should be similar, since, generally, the cases
all
> have the same init times).
>
> B) the matches for other models (ex: GFSO, etc.) include storms like
> Florence (2012) which is an example of a non-RI case (it was a weak
> tropical storm that dissipated within a few days).
>
> Am I doing something incorrect when setting this flag? The number of
storms
> that DID undergo RI during our sampling period was relatively small,
so it
> may be a moot point, but this behavior seems odd.
>
> Cheers,
>
> -Colin
>
> --
> Colin M. Zarzycki, ASP postdoctoral fellow
> Atmospheric Modeling and Predictability
> National Center for Atmospheric Research
> http://www.colinzarzycki.com
>
>
--
Colin M. Zarzycki, ASP postdoctoral fellow
Atmospheric Modeling and Predictability
National Center for Atmospheric Research
http://www.colinzarzycki.com
------------------------------------------------
Subject: rapid_intens filtering questions with tc_stat in MET-TC
From: John Halley Gotway
Time: Tue Sep 16 15:34:15 2014
Colin,
Thanks for the heads up about the documentation. You're right, we
missed
that while updating the rapid_inten documentation for the met-5.0
release.
We'll update it in the working version and include it in the next
release.
We did generalize the rapid intensification/weakening logic for met-
5.0 to
make it more configurable. Here's a selection from the file
met-5.0/data/config/README_TC...
//
// Specify whether only those track points for which rapid
intensification
// or weakening of the maximum wind speed occurred in the previous
time
// step should be retained.
// The NHC considers a 24-hour change >=30 kts to constitute rapid
// intensification or weakening.
// May modify using the "-rapid_inten_track", "-rapid_inten_time",
// "-rapid_inten_exact", and "rapid_inten_thresh" job command options.
//
rapid_inten = {
track = NONE; // Specify which track types to search (NONE,
ADECK, BDECK, or BOTH)
time = 24; // Rapid intensification/weakening time period in
HHMMSS format.
exact = TRUE; // Use the exact or maximum intensity difference
over the time period.
thresh = >=30.0; // Threshold for the intensity change.
}
Please let me know if you have additional questions about defining
RI/RW
criteria.
Thanks,
John Halley Gotway
met_help at ucar.edu
On Tue, Sep 16, 2014 at 12:06 PM, Colin Zarzycki via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=68989 >
>
> It appears many of my issues are mitigated by moving to MET 5.0. The
> results are at least somewhat consistent between the models
> using -rapid_inten_track BDECK (which I believe is the behavior I
want).
>
> Of note, the 5.0 documentation does not document the command line
> implementation of the new RI settings (I had to peek at the source
code to
> note that each flag (-rapid_inten_track, rapid_intens_thresh, etc.)
needed
> to be specified as opposed to one rapid_inten setting).
>
> It still remains unclear to me how RI is defined, but I will play
around
> with it a bit first.
>
> -Colin
>
> On Tue, Sep 16, 2014 at 10:28 AM, met_help at ucar.edu via RT <
> met_help at ucar.edu> wrote:
>
> > Greetings,
> >
> > This message has been automatically generated in response to the
> > creation of a trouble ticket regarding:
> > "rapid_intens filtering questions with tc_stat in MET-TC",
> > a summary of which appears below.
> >
> > There is no need to reply to this message right now. Your ticket
has
> been
> > assigned an ID of [rt.rap.ucar.edu #68989].
> >
> > Please include the string:
> >
> > [rt.rap.ucar.edu #68989]
> >
> > in the subject line of all future correspondence about this issue.
To do
> > so,
> > you may reply to this message.
> >
> > Thank you,
> > met_help at ucar.edu
> >
> >
-------------------------------------------------------------------------
> > Hi,
> >
> > I was wondering if I could get clarification on how applying the
> > "rapid_intens" flag in tc_stat functions.
> >
> > I was under the impression that it would find periods of rapid
> > intensification in the b-deck, and then only select forecast times
that
> > applied to the previous 24 hours (or perhaps earlier).
> >
> > EX: Hurricane John undergoes rapid intensification (RI) from 00Z
on the
> > 21st to 00Z on the 22nd.
> > in the b-deck, WIND(00Z 22nd) - WIND(00Z 21st) >= 30 kts, so RI
achieved.
> > Therefore, tc_stat should keep all forecasts from 00Z on the 21st
(and
> > perhaps 00Z on the 20th and 00Z on the 19th, etc.) depending on
whether
> you
> > want the forecast from where RI began or all forecasts which
include lead
> > times where the RI occured.
> >
> > However, if I apply the rapid_intens criteria (on Yellowstone in
> > /glade/p/work/zarzycki/test_met/) when just testing the one of the
> provided
> > R scripts and some random models...
> >
> > Rscript
/glade/u/home/zarzycki/METv4.1/scripts/Rscripts/plot_tcmpr.R
> > -lookin /glade/u/home/zarzycki/work/test_met/out_tcmpr.tcst
-filter
> > "-amodel EGRR,GFDL,CMZ,GFSO,HWRF -rapid_inten true
-rapid_inten_thresh
> > ge30" -save_data "./save_the_data.out" -dep "TK_ERR,AMAX_WIND-
BMAX_WIND"
> > -series AMODEL EGRR,GFDL,CMZ,GFSO,HWRF -plot MEAN,BOXPLOT -outdir
> > /glade/u/home/zarzycki/work/test_met/ -lead
> 12,24,48,72,96,120,144,168,192
> > -no_ee
> >
> > I get A) a tremendously large number of matches for our model and
not
> other
> > models at 24 hour lead, which shouldn't be the case if we are only
using
> > b-deck (these numbers should be similar, since, generally, the
cases all
> > have the same init times).
> >
> > B) the matches for other models (ex: GFSO, etc.) include storms
like
> > Florence (2012) which is an example of a non-RI case (it was a
weak
> > tropical storm that dissipated within a few days).
> >
> > Am I doing something incorrect when setting this flag? The number
of
> storms
> > that DID undergo RI during our sampling period was relatively
small, so
> it
> > may be a moot point, but this behavior seems odd.
> >
> > Cheers,
> >
> > -Colin
> >
> > --
> > Colin M. Zarzycki, ASP postdoctoral fellow
> > Atmospheric Modeling and Predictability
> > National Center for Atmospheric Research
> > http://www.colinzarzycki.com
> >
> >
>
>
> --
> Colin M. Zarzycki, ASP postdoctoral fellow
> Atmospheric Modeling and Predictability
> National Center for Atmospheric Research
> http://www.colinzarzycki.com
>
>
------------------------------------------------
More information about the Met_help
mailing list