[Met_help] [rt.rap.ucar.edu #70294] History for nam rainfall forecast
John Halley Gotway via RT
met_help at ucar.edu
Fri Feb 20 09:52:48 MST 2015
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Dear MET help,
I want to produce NAM rainfall forecast skill score with MET (comparing with stage IV observation).
Is there a way of doing so?
NAM data are downloaded from here:
ftp://ftp.ncep.noaa.gov/pub/data/nccf/com/nam/prod/
They are in GRIB format, however, in order to get their three hourly rainfall, we need an additional post processing step.
Do you know if there are any softwares that we can use to preprocess NAM for MET?
Thanks,
Di
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: nam rainfall forecast
From: John Halley Gotway
Time: Wed Jan 07 13:34:33 2015
Hello Di,
You should definitely be able to use MET to compare NAM to StageIV
observations. Since you're comparing a gridded forecast to a gridded
observation, you'd use the MET grid_stat tool. Here are the steps I'd
suggest:
(1) Decide the grid over which you'd like to compute that statistics -
either NAM grid or the 4km StageIV grid.
(2) Decide what rainfall accumulation interval you'd like to evaluate
-
sound like you're interested in 3-hourly accumulations.
(3) Use copygb to either regrid the NAM to the StageIV domain or vice
versa. Examples of running copygb are included in the MET online
tutorial:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/copygb/index.php
(4) Now that you have the forecast and observations on a common grid,
run
the MET pcp_combine tool to put the precipitation into 3-hourly bins.
StageIV is available in 1, 6, and 24 hourly chunks. To get 3 hourly,
you
could add up 3 1-hourly files (e.g. pcp_combine -add). NAM
precipitation
may already be available in 3-hourly chunks. If so, no additional
work is
required. If not, you may need to run pcp_combine to get it into 3-
hourly
bins.
(5) Copy over and setup a configuration file for the grid_stat tool.
Specify the geographic regions over which you want to compute
statistics,
the precipitation thresholds you'd like to use, and the statistical
output
types you'd like to see (usually contingency table counts (CTC) and
contingency table statistics (CTS), but maybe continuous statistics
(CNT)
too).
(6) Run grid_stat once for each 3-hourly accumulation you'd like to
evaluate.
I'd suggest putting these steps together into a script that you'd run
once
for each forecast to be evaluated.
As for the statistics MET computes, please refer to appendix C of the
MET
user's guide:
http://www.dtcenter.org/met/users/docs/overview.php
As for skill scores, the CTS line type includes the Gilbert Skill
Score and
Heidke Skill Score, along with several other statistics.
Hope that helps get you going.
Thanks,
John Halley Gotway
met_help at ucar.edu
On Wed, Jan 7, 2015 at 12:25 PM, Wu, Di[SCIENCE SYSTEMS APPLICATIONS]
via
RT <met_help at ucar.edu> wrote:
>
> Wed Jan 07 12:25:43 2015: Request 70294 was acted upon.
> Transaction: Ticket created by di.wu at nasa.gov
> Queue: met_help
> Subject: nam rainfall forecast
> Owner: Nobody
> Requestors: di.wu at nasa.gov
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=70294 >
>
>
> Dear MET help,
>
> I want to produce NAM rainfall forecast skill score with MET
(comparing
> with stage IV observation).
> Is there a way of doing so?
> NAM data are downloaded from here:
> ftp://ftp.ncep.noaa.gov/pub/data/nccf/com/nam/prod/
> They are in GRIB format, however, in order to get their three hourly
> rainfall, we need an additional post processing step.
> Do you know if there are any softwares that we can use to preprocess
NAM
> for MET?
>
> Thanks,
> Di
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #70294] nam rainfall forecast
From: Wu, Di[SCIENCE SYSTEMS APPLICATIONS]
Time: Tue Feb 10 13:56:40 2015
Hi John,
Thanks for your reply, it helps to familiarize me with the UPP and MET
procedures.
However, there are still problems before proceeding to step 1.
NAM data are not automatically 3 hours accumulation.
00UTC 0 accumulation
03UTC is 0-3 hour accumulation
06UTC is 0-6 hour accumulation
09UTC is 0-9 hour accumulation
12UTC is 0-12 hour accumulation
15UTC is 12-15 hour accumulation
18UTC is 12-18 hour accumulation
...
So can UPP take care of this problem and get 3 hours accumulation?
Thanks,
Di
On Jan 7, 2015, at 3:34 PM, John Halley Gotway via RT wrote:
> Hello Di,
>
> You should definitely be able to use MET to compare NAM to StageIV
> observations. Since you're comparing a gridded forecast to a
gridded
> observation, you'd use the MET grid_stat tool. Here are the steps
I'd
> suggest:
> (1) Decide the grid over which you'd like to compute that statistics
-
> either NAM grid or the 4km StageIV grid.
> (2) Decide what rainfall accumulation interval you'd like to
evaluate -
> sound like you're interested in 3-hourly accumulations.
> (3) Use copygb to either regrid the NAM to the StageIV domain or
vice
> versa. Examples of running copygb are included in the MET online
tutorial:
>
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/copygb/index.php
> (4) Now that you have the forecast and observations on a common
grid, run
> the MET pcp_combine tool to put the precipitation into 3-hourly
bins.
> StageIV is available in 1, 6, and 24 hourly chunks. To get 3
hourly, you
> could add up 3 1-hourly files (e.g. pcp_combine -add). NAM
precipitation
> may already be available in 3-hourly chunks. If so, no additional
work is
> required. If not, you may need to run pcp_combine to get it into 3-
hourly
> bins.
> (5) Copy over and setup a configuration file for the grid_stat tool.
> Specify the geographic regions over which you want to compute
statistics,
> the precipitation thresholds you'd like to use, and the statistical
output
> types you'd like to see (usually contingency table counts (CTC) and
> contingency table statistics (CTS), but maybe continuous statistics
(CNT)
> too).
> (6) Run grid_stat once for each 3-hourly accumulation you'd like to
> evaluate.
>
> I'd suggest putting these steps together into a script that you'd
run once
> for each forecast to be evaluated.
>
> As for the statistics MET computes, please refer to appendix C of
the MET
> user's guide:
> http://www.dtcenter.org/met/users/docs/overview.php
>
> As for skill scores, the CTS line type includes the Gilbert Skill
Score and
> Heidke Skill Score, along with several other statistics.
>
> Hope that helps get you going.
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
> On Wed, Jan 7, 2015 at 12:25 PM, Wu, Di[SCIENCE SYSTEMS
APPLICATIONS] via
> RT <met_help at ucar.edu> wrote:
>
>>
>> Wed Jan 07 12:25:43 2015: Request 70294 was acted upon.
>> Transaction: Ticket created by di.wu at nasa.gov
>> Queue: met_help
>> Subject: nam rainfall forecast
>> Owner: Nobody
>> Requestors: di.wu at nasa.gov
>> Status: new
>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=70294 >
>>
>>
>> Dear MET help,
>>
>> I want to produce NAM rainfall forecast skill score with MET
(comparing
>> with stage IV observation).
>> Is there a way of doing so?
>> NAM data are downloaded from here:
>> ftp://ftp.ncep.noaa.gov/pub/data/nccf/com/nam/prod/
>> They are in GRIB format, however, in order to get their three
hourly
>> rainfall, we need an additional post processing step.
>> Do you know if there are any softwares that we can use to
preprocess NAM
>> for MET?
>>
>> Thanks,
>> Di
>>
>>
>
------------------------------------------------
Subject: nam rainfall forecast
From: John Halley Gotway
Time: Wed Feb 11 10:07:36 2015
Di,
Here's what you should do...
(1) Run copygb to either put your GRIB model output onto the Stave IV
domain or put the GRIB Stage IV data onto your model domain.
(2) Run the MET pcp_combine tool to put the NAM data into 3-hourly
accumulations. Using the information you sent, you'd do it as
follows:
0-3 hour accumulation is already in the 03UTC file. I'd suggest
running
this file through pcp_combine as a pass-through to put it into NetCDF
format:
pcp_combine -add 03_file.grb 03 APCP_00_03.nc
3-6 hour accumulation. Subtract 0-6 and 0-3 accumulations:
pcp_combine -subtract 06_file.grb 06 03_file.grb 03 APCP_03_06.nc
6-9 hour accumulation. Subtract 0-9 and 0-6 accumulations:
pcp_combine -subtract 09_file.grb 09 06_file.grb 06 APCP_06_09.nc
9-12 hour accumulation. Subtract 0-12 and 0-9 accumulations:
pcp_combine -subtract 12_file.grb 12 09_file.grb 09 APCP_09_12.nc
12-15 hour accumulation. Just run as a pass-through again:
pcp_combine -add 15_file.grb 03 APCP_12_15.nc
15-18 hour accumulation. Subtract 12-18 and 12-15 accumulations:
pcp_combine -subtract 18_file.grb 06 15_file.grb 03 APCP_15_18.nc
And so on...
I suggest running 0-3 and 12-15 through pcp_combine even though they
already have the 3-hour accumulation. That way, all of your NAM files
will
be in the same file format, and you can use the same configuration
file
settings for the other MET tools (grid_stat, mode, etc.). If the NAM
files
are a mix of GRIB and NetCDF, the logic would need to be a bit more
complicated.
Does that all make sense?
Thanks,
John
On Tue, Feb 10, 2015 at 1:56 PM, Wu, Di[SCIENCE SYSTEMS APPLICATIONS]
via
RT <met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=70294 >
>
> Hi John,
>
> Thanks for your reply, it helps to familiarize me with the UPP and
MET
> procedures.
> However, there are still problems before proceeding to step 1.
> NAM data are not automatically 3 hours accumulation.
> 00UTC 0 accumulation
> 03UTC is 0-3 hour accumulation
> 06UTC is 0-6 hour accumulation
> 09UTC is 0-9 hour accumulation
> 12UTC is 0-12 hour accumulation
>
> 15UTC is 12-15 hour accumulation
> 18UTC is 12-18 hour accumulation
> ...
>
> So can UPP take care of this problem and get 3 hours accumulation?
>
> Thanks,
> Di
>
>
>
> On Jan 7, 2015, at 3:34 PM, John Halley Gotway via RT wrote:
>
> > Hello Di,
> >
> > You should definitely be able to use MET to compare NAM to StageIV
> > observations. Since you're comparing a gridded forecast to a
gridded
> > observation, you'd use the MET grid_stat tool. Here are the steps
I'd
> > suggest:
> > (1) Decide the grid over which you'd like to compute that
statistics -
> > either NAM grid or the 4km StageIV grid.
> > (2) Decide what rainfall accumulation interval you'd like to
evaluate -
> > sound like you're interested in 3-hourly accumulations.
> > (3) Use copygb to either regrid the NAM to the StageIV domain or
vice
> > versa. Examples of running copygb are included in the MET online
> tutorial:
> >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/copygb/index.php
> > (4) Now that you have the forecast and observations on a common
grid, run
> > the MET pcp_combine tool to put the precipitation into 3-hourly
bins.
> > StageIV is available in 1, 6, and 24 hourly chunks. To get 3
hourly, you
> > could add up 3 1-hourly files (e.g. pcp_combine -add). NAM
precipitation
> > may already be available in 3-hourly chunks. If so, no additional
work
> is
> > required. If not, you may need to run pcp_combine to get it into
> 3-hourly
> > bins.
> > (5) Copy over and setup a configuration file for the grid_stat
tool.
> > Specify the geographic regions over which you want to compute
statistics,
> > the precipitation thresholds you'd like to use, and the
statistical
> output
> > types you'd like to see (usually contingency table counts (CTC)
and
> > contingency table statistics (CTS), but maybe continuous
statistics (CNT)
> > too).
> > (6) Run grid_stat once for each 3-hourly accumulation you'd like
to
> > evaluate.
> >
> > I'd suggest putting these steps together into a script that you'd
run
> once
> > for each forecast to be evaluated.
> >
> > As for the statistics MET computes, please refer to appendix C of
the MET
> > user's guide:
> > http://www.dtcenter.org/met/users/docs/overview.php
> >
> > As for skill scores, the CTS line type includes the Gilbert Skill
Score
> and
> > Heidke Skill Score, along with several other statistics.
> >
> > Hope that helps get you going.
> >
> > Thanks,
> > John Halley Gotway
> > met_help at ucar.edu
> >
> > On Wed, Jan 7, 2015 at 12:25 PM, Wu, Di[SCIENCE SYSTEMS
APPLICATIONS] via
> > RT <met_help at ucar.edu> wrote:
> >
> >>
> >> Wed Jan 07 12:25:43 2015: Request 70294 was acted upon.
> >> Transaction: Ticket created by di.wu at nasa.gov
> >> Queue: met_help
> >> Subject: nam rainfall forecast
> >> Owner: Nobody
> >> Requestors: di.wu at nasa.gov
> >> Status: new
> >> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=70294 >
> >>
> >>
> >> Dear MET help,
> >>
> >> I want to produce NAM rainfall forecast skill score with MET
(comparing
> >> with stage IV observation).
> >> Is there a way of doing so?
> >> NAM data are downloaded from here:
> >> ftp://ftp.ncep.noaa.gov/pub/data/nccf/com/nam/prod/
> >> They are in GRIB format, however, in order to get their three
hourly
> >> rainfall, we need an additional post processing step.
> >> Do you know if there are any softwares that we can use to
preprocess NAM
> >> for MET?
> >>
> >> Thanks,
> >> Di
> >>
> >>
> >
>
>
>
>
------------------------------------------------
More information about the Met_help
mailing list