[Met_help] [rt.rap.ucar.edu #82839] History for Question regarding the WRF-precipitation and NCEP Stage IV evaluation
John Halley Gotway via RT
met_help at ucar.edu
Tue Jul 9 12:04:09 MDT 2019
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi,
Here I have a quick question regarding how to use Grid-STAT to evaluate the precipitation from WRF with the observation NCEP Stage IV. The precipitation of WRF is the sum of field "RAINNC" and "RAINC", there are accumulated precipitation from the start time of the model, I have 10 days (240 hours, I would like to use the last hour). However, the NCEP Stage IV is the hourly precipitation in the format (GRIB). My plan is to use PCP-Combine to sum the precipitations of all the NCEP Stage IV as the accumulated precipitation and use "Grid-STAT" and "Stat-Analysis" to have the grid-comparison.
However, I got stuck with the following questions:
1. How to re-grid both the results of the last hour of WRF and the sum of NCEP Stage IV?
2. How to compare the sum of fields "RAINNC" and "RAINC" of WRF with NCEP Stage IV?
3. Is there any template or tutorial for comparing the precipitation of WRF with NCEP Stage IV?
I look forward to hearing from you!
Sincerely,
Zhonghua
---------------------------------------------------
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Question regarding the WRF-precipitation and NCEP Stage IV evaluation
From: John Halley Gotway
Time: Thu Nov 02 15:46:26 2017
Zhonghua,
I see you'd like to use MET to compare the NetCDF output of WRF to
NCEP
StageIV.
First, MET cannot read the raw NetCDF output of WRF directly. It must
first be pre-processed using the wrf_interp utility or by running the
Unified PostProcessor. I'd strongly recommend running UPP which
writes
it's output in either GRIB1 or GRIB2, both of which MET can handle
well.
The UPP output will contain a field of accumulated precipitation
(RAINC +
RAINNC).
If you'd like to evaluation hourly precip against StageIV, you'd run
MET's
pcp_combine tool using the -subtract option. Let's say your UPP
output
contains a runtime accumulation of precip in files name wrfprs_##.grb
where
## is the forecast hour. You'd run pcp_combine like this:
pcp_combine -subtract wrfprs_24.grb 24 wrfprs_23.grb 23
apcp_23_to_24.nc
That'd get the 24 hours of precip from the first file and subtract off
the
23 hours of precip from the 2nd file. That difference is the precip
that
fell between hours 23 and 24.
If you were using the NetCDF output of wrf_interp, you'd need to run
pcp_combine 3 times:
(1) Add RAINC + RAINNC for hour 23.
(2) Add RAINC + RAINNC for hour 24.
(3) Subtract 24 minus 23.
Let's assume you have you precip in matching time accumulation
intervals
for the forecast and observation. Putting them on the same grid is
pretty
easy. The Grid-Stat configuration file has a section named "regrid"
where
you specify the evaluation domain. Regridding to the observation
(i.e.
StageIV) domain is as easy as saying "to_grid = OBS;".
Here's a link to the MET User's Guide:
https://dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v6.0.pdf
And here's a link to the online tutorial which contains some examples
of
running pcp_combine and regridding:
https://dtcenter.org/met/users/support/online_tutorial/METv6.0/index.php
Hope that helps get you going.
Thanks,
John Halley Gotway
On Thu, Nov 2, 2017 at 9:51 AM, Zheng, Zhonghua via RT
<met_help at ucar.edu>
wrote:
>
> Thu Nov 02 09:51:42 2017: Request 82839 was acted upon.
> Transaction: Ticket created by zzheng25 at illinois.edu
> Queue: met_help
> Subject: Question regarding the WRF-precipitation and NCEP
Stage IV
> evaluation
> Owner: Nobody
> Requestors: zzheng25 at illinois.edu
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82839 >
>
>
> Hi,
>
> Here I have a quick question regarding how to use Grid-STAT to
evaluate
> the precipitation from WRF with the observation NCEP Stage IV. The
> precipitation of WRF is the sum of field "RAINNC" and "RAINC", there
are
> accumulated precipitation from the start time of the model, I have
10 days
> (240 hours, I would like to use the last hour). However, the NCEP
Stage IV
> is the hourly precipitation in the format (GRIB). My plan is to use
> PCP-Combine to sum the precipitations of all the NCEP Stage IV as
the
> accumulated precipitation and use "Grid-STAT" and "Stat-Analysis" to
have
> the grid-comparison.
> However, I got stuck with the following questions:
> 1. How to re-grid both the results of the last hour of WRF and the
sum of
> NCEP Stage IV?
> 2. How to compare the sum of fields "RAINNC" and "RAINC" of WRF with
NCEP
> Stage IV?
> 3. Is there any template or tutorial for comparing the precipitation
of
> WRF with NCEP Stage IV?
>
> I look forward to hearing from you!
>
> Sincerely,
> Zhonghua
> ---------------------------------------------------
>
>
>
>
------------------------------------------------
Subject: Question regarding the WRF-precipitation and NCEP Stage IV evaluation
From: Zheng, Zhonghua
Time: Mon Nov 06 14:54:50 2017
Hi John,
Thank you very much for your help! I also get confused about the
configuration file and how to have the multiple analysis.
For the configuration file, is there any template for the
precipitation comparison specific to WRF and NCEP Stage IV?
For the Grid-Stat, I see the command only has two files comparison,
how to have a series of comparison of the 240 hours comparison (240
files from WRF-UPP and 240 files from NCEP Stage IV) and aggregate
results?
I look forward to hearing from you!
Sincerely,
Zhonghua
---------------------------------------------------
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Thursday, November 02, 2017 4:46 PM
To: Zheng, Zhonghua
Subject: Re: [rt.rap.ucar.edu #82839] Question regarding the WRF-
precipitation and NCEP Stage IV evaluation
Zhonghua,
I see you'd like to use MET to compare the NetCDF output of WRF to
NCEP
StageIV.
First, MET cannot read the raw NetCDF output of WRF directly. It must
first be pre-processed using the wrf_interp utility or by running the
Unified PostProcessor. I'd strongly recommend running UPP which writes
it's output in either GRIB1 or GRIB2, both of which MET can handle
well.
The UPP output will contain a field of accumulated precipitation
(RAINC +
RAINNC).
If you'd like to evaluation hourly precip against StageIV, you'd run
MET's
pcp_combine tool using the -subtract option. Let's say your UPP output
contains a runtime accumulation of precip in files name wrfprs_##.grb
where
## is the forecast hour. You'd run pcp_combine like this:
pcp_combine -subtract wrfprs_24.grb 24 wrfprs_23.grb 23
apcp_23_to_24.nc
That'd get the 24 hours of precip from the first file and subtract off
the
23 hours of precip from the 2nd file. That difference is the precip
that
fell between hours 23 and 24.
If you were using the NetCDF output of wrf_interp, you'd need to run
pcp_combine 3 times:
(1) Add RAINC + RAINNC for hour 23.
(2) Add RAINC + RAINNC for hour 24.
(3) Subtract 24 minus 23.
Let's assume you have you precip in matching time accumulation
intervals
for the forecast and observation. Putting them on the same grid is
pretty
easy. The Grid-Stat configuration file has a section named "regrid"
where
you specify the evaluation domain. Regridding to the observation (i.e.
StageIV) domain is as easy as saying "to_grid = OBS;".
Here's a link to the MET User's Guide:
https://dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v6.0.pdf
And here's a link to the online tutorial which contains some examples
of
running pcp_combine and regridding:
https://dtcenter.org/met/users/support/online_tutorial/METv6.0/index.php
Hope that helps get you going.
Thanks,
John Halley Gotway
On Thu, Nov 2, 2017 at 9:51 AM, Zheng, Zhonghua via RT
<met_help at ucar.edu>
wrote:
>
> Thu Nov 02 09:51:42 2017: Request 82839 was acted upon.
> Transaction: Ticket created by zzheng25 at illinois.edu
> Queue: met_help
> Subject: Question regarding the WRF-precipitation and NCEP Stage IV
> evaluation
> Owner: Nobody
> Requestors: zzheng25 at illinois.edu
> Status: new
> Ticket <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82839
>
>
>
> Hi,
>
> Here I have a quick question regarding how to use Grid-STAT to
evaluate
> the precipitation from WRF with the observation NCEP Stage IV. The
> precipitation of WRF is the sum of field "RAINNC" and "RAINC", there
are
> accumulated precipitation from the start time of the model, I have
10 days
> (240 hours, I would like to use the last hour). However, the NCEP
Stage IV
> is the hourly precipitation in the format (GRIB). My plan is to use
> PCP-Combine to sum the precipitations of all the NCEP Stage IV as
the
> accumulated precipitation and use "Grid-STAT" and "Stat-Analysis" to
have
> the grid-comparison.
> However, I got stuck with the following questions:
> 1. How to re-grid both the results of the last hour of WRF and the
sum of
> NCEP Stage IV?
> 2. How to compare the sum of fields "RAINNC" and "RAINC" of WRF with
NCEP
> Stage IV?
> 3. Is there any template or tutorial for comparing the precipitation
of
> WRF with NCEP Stage IV?
>
> I look forward to hearing from you!
>
> Sincerely,
> Zhonghua
> ---------------------------------------------------
>
>
>
>
------------------------------------------------
Subject: Question regarding the WRF-precipitation and NCEP Stage IV evaluation
From: John Halley Gotway
Time: Mon Nov 06 15:25:55 2017
Zhonghua,
The Grid-Stat tool is designed to be run once for each output time.
So
you'd use a shell script (or python) to loop over your model output
times
and run grid-stat for each time. Then, if you'd like to aggregate
results
through time, you'd run the stat_analysis tool to do so.
Examples of these can be found in the MET online tutorial:
https://dtcenter.org/met/users/support/online_tutorial/METv6.0/index.php
The Grid-Stat tool computes statistics computed over spatial areas.
However, the Series-Analysis tool can be run to compute statistics
separately for each model grid point *through* time. That's helpful
to see
how model performance varies over your domain.
Both the spatial summary statistics from Grid-Stat and the time
summary
statistics at each grid point from Series-Analysis are useful. They
just
answer different questions about your model performance.
John
On Mon, Nov 6, 2017 at 2:54 PM, Zheng, Zhonghua via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82839 >
>
> Hi John,
>
> Thank you very much for your help! I also get confused about the
> configuration file and how to have the multiple analysis.
> For the configuration file, is there any template for the
precipitation
> comparison specific to WRF and NCEP Stage IV?
> For the Grid-Stat, I see the command only has two files comparison,
how to
> have a series of comparison of the 240 hours comparison (240 files
from
> WRF-UPP and 240 files from NCEP Stage IV) and aggregate results?
>
> I look forward to hearing from you!
>
> Sincerely,
> Zhonghua
> ---------------------------------------------------
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 02, 2017 4:46 PM
> To: Zheng, Zhonghua
> Subject: Re: [rt.rap.ucar.edu #82839] Question regarding the
> WRF-precipitation and NCEP Stage IV evaluation
>
> Zhonghua,
>
> I see you'd like to use MET to compare the NetCDF output of WRF to
NCEP
> StageIV.
>
> First, MET cannot read the raw NetCDF output of WRF directly. It
must
> first be pre-processed using the wrf_interp utility or by running
the
> Unified PostProcessor. I'd strongly recommend running UPP which
writes
> it's output in either GRIB1 or GRIB2, both of which MET can handle
well.
> The UPP output will contain a field of accumulated precipitation
(RAINC +
> RAINNC).
>
> If you'd like to evaluation hourly precip against StageIV, you'd run
MET's
> pcp_combine tool using the -subtract option. Let's say your UPP
output
> contains a runtime accumulation of precip in files name
wrfprs_##.grb where
> ## is the forecast hour. You'd run pcp_combine like this:
>
> pcp_combine -subtract wrfprs_24.grb 24 wrfprs_23.grb 23
apcp_23_to_24.nc
>
> That'd get the 24 hours of precip from the first file and subtract
off the
> 23 hours of precip from the 2nd file. That difference is the precip
that
> fell between hours 23 and 24.
>
> If you were using the NetCDF output of wrf_interp, you'd need to run
> pcp_combine 3 times:
> (1) Add RAINC + RAINNC for hour 23.
> (2) Add RAINC + RAINNC for hour 24.
> (3) Subtract 24 minus 23.
>
> Let's assume you have you precip in matching time accumulation
intervals
> for the forecast and observation. Putting them on the same grid is
pretty
> easy. The Grid-Stat configuration file has a section named "regrid"
where
> you specify the evaluation domain. Regridding to the observation
(i.e.
> StageIV) domain is as easy as saying "to_grid = OBS;".
>
> Here's a link to the MET User's Guide:
>
https://dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v6.0.pdf
>
> And here's a link to the online tutorial which contains some
examples of
> running pcp_combine and regridding:
>
https://dtcenter.org/met/users/support/online_tutorial/METv6.0/index.php
>
> Hope that helps get you going.
>
> Thanks,
> John Halley Gotway
>
> On Thu, Nov 2, 2017 at 9:51 AM, Zheng, Zhonghua via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > Thu Nov 02 09:51:42 2017: Request 82839 was acted upon.
> > Transaction: Ticket created by zzheng25 at illinois.edu
> > Queue: met_help
> > Subject: Question regarding the WRF-precipitation and NCEP Stage
IV
> > evaluation
> > Owner: Nobody
> > Requestors: zzheng25 at illinois.edu
> > Status: new
> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82839 >
> >
> >
> > Hi,
> >
> > Here I have a quick question regarding how to use Grid-STAT to
evaluate
> > the precipitation from WRF with the observation NCEP Stage IV. The
> > precipitation of WRF is the sum of field "RAINNC" and "RAINC",
there are
> > accumulated precipitation from the start time of the model, I have
10
> days
> > (240 hours, I would like to use the last hour). However, the NCEP
Stage
> IV
> > is the hourly precipitation in the format (GRIB). My plan is to
use
> > PCP-Combine to sum the precipitations of all the NCEP Stage IV as
the
> > accumulated precipitation and use "Grid-STAT" and "Stat-Analysis"
to have
> > the grid-comparison.
> > However, I got stuck with the following questions:
> > 1. How to re-grid both the results of the last hour of WRF and the
sum of
> > NCEP Stage IV?
> > 2. How to compare the sum of fields "RAINNC" and "RAINC" of WRF
with NCEP
> > Stage IV?
> > 3. Is there any template or tutorial for comparing the
precipitation of
> > WRF with NCEP Stage IV?
> >
> > I look forward to hearing from you!
> >
> > Sincerely,
> > Zhonghua
> > ---------------------------------------------------
> >
> >
> >
> >
>
>
>
------------------------------------------------
More information about the Met_help
mailing list