[Met_help] [rt.rap.ucar.edu #88674] History for Question on ensemble_stat when using point_obs

Mon Jan 28 16:21:54 MST 2019

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

Hi,

I am trying to run ensemble_stat on point observations for Total
Precipitation (APCP).

Even though the met config file I am using does not have any A03 level in
it, the met_log shows it is looking for

APCP/A03 in the GEPS ensemble GFS forecast file. Then I get the following
WARNING message:

WARNING: MetGrib2DataFile::data_plane() - No matching record found for
'APCP/A03'

WARNING: 

WARNING: 

WARNING: process_n_vld() -> ensemble field "APCP/A03" not found in file
"/home/sarasammas/geps_verify/reduced_gribs/2018120500/GFS/GFS_E00_201812050
00000_f012.gr2"

For the run as shown in the attached log file, I was using a non-standard
grib_table, thinking the observation file was created using that grib_table.

I had no trouble running point_stat on the same pair of forecast file and
observation file.

For your reference I am attaching the log file, my config file and the
observation file.

Thanks

Suseela

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

Subject: Question on ensemble_stat when using point_obs
From: John Halley Gotway
Time: Mon Jan 28 14:54:40 2019

Suseela,

I see you have a question about your Ensemble-Stat config file.
Thanks for
sending an example.

You make no reference to 3-hourly precip in your Ensemble-Stat config
file,
and yet, Ensemble-Stat is trying to process that data.

Here's what's going on... When you run Ensemble-Stat, you pass to it a
configuration file on the command line.  The one you sent me is named
"ensemble_stat_model_obs_apcp.cfg".  At runtime, Ensemble-Stat (like
all
the other MET tools with config files) actually reads multiple config
files:
 (1) Constants file: met-8.0/share/met/ConfigConstants
 (2) Default for the tool: met-
8.0/share/met/EnsembleStatConfig_default
 (3) User-specified one from the command
line:ensemble_stat_model_obs_apcp.cfg

These reads are cumulative.  So the settings you put in your user
config
file override the defaults.  Anything you do not override remains at
it's
default setting.  And that's what's happening here.  Your config file
is
missing the ensemble (ens) section which lists the fields for which
you
want ensemble means and relative frequencies computed.  So it's using
the
default setting which has 3-hourly accumulated precip
(met-8.0/share/met/EnsembleStatConfig_default).

Frankly an empty list probably is a better default value... but it's a
balance between good defaults and good examples.

Please try adding the ens section into your config file and setting it
to
an empty list:

*ens = {   ens_thresh = 1.0;   vld_thresh = 1.0;   field = [ ];}*

Hope that helps.  Note that I did not go through your config file
carefully
to see if other items are missing.  So it'd be a good idea to compare
yours
with the default to think about your config file choices.

Hope that helps.

Thanks,
John Halley Gotway

On Mon, Jan 28, 2019 at 2:38 PM SARASAMMA, SUSEELA T CTR USAF AFMC
AFLCMC/HBAW-OL via RT <met_help at ucar.edu> wrote:

>
> Mon Jan 28 14:38:15 2019: Request 88674 was acted upon.
> Transaction: Ticket created by suseela.sarasamma.ctr at us.af.mil
>        Queue: met_help
>      Subject: Question on ensemble_stat when using point_obs
>        Owner: Nobody
>   Requestors: suseela.sarasamma.ctr at us.af.mil
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=88674 >
>
>
> Hi,
>
> I am trying to run ensemble_stat on point observations for Total
> Precipitation (APCP).
>
>
>
> Even though the met config file I am using does not have any A03
level in
> it, the met_log shows it is looking for
>
> APCP/A03 in the GEPS ensemble GFS forecast file. Then I get the
following
> WARNING message:
>
>
>
> WARNING: MetGrib2DataFile::data_plane() - No matching record found
for
> 'APCP/A03'
>
> WARNING:
>
> WARNING:
>
> WARNING: process_n_vld() -> ensemble field "APCP/A03" not found in
file
>
>
"/home/sarasammas/geps_verify/reduced_gribs/2018120500/GFS/GFS_E00_201812050
> 00000_f012.gr2"
>
>
>
> For the run as shown in the attached log file, I was using a non-
standard
> grib_table, thinking the observation file was created using that
> grib_table.
>
>
>
> I had no trouble running point_stat on the same pair of forecast
file and
> observation file.
>
>
>
> For your reference I am attaching the log file, my config file and
the
> observation file.
>
>
>
> Thanks
>
> Suseela
>
>
>
>
>

------------------------------------------------
Subject: RE: [Non-DoD Source] Re: [rt.rap.ucar.edu #88674] Question on ensemble_stat when using point_obs
From: SARASAMMA, SUSEELA T CTR USAF AFMC AFLCMC/HBAW-OL
Time: Mon Jan 28 15:13:13 2019

John,
Thank you so much for the quick response. You found the problem in my
config file.
It works now.
suseela

-----Original Message-----
From: John Halley Gotway via RT <met_help at ucar.edu>
Sent: Monday, January 28, 2019 3:55 PM
To: SARASAMMA, SUSEELA T CTR USAF AFMC AFLCMC/HBAW-OL
<suseela.sarasamma.ctr at us.af.mil>
Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #88674] Question on
ensemble_stat when using point_obs

Suseela,

I see you have a question about your Ensemble-Stat config file.
Thanks for sending an example.

You make no reference to 3-hourly precip in your Ensemble-Stat config
file, and yet, Ensemble-Stat is trying to process that data.

Here's what's going on... When you run Ensemble-Stat, you pass to it a
configuration file on the command line.  The one you sent me is named
"ensemble_stat_model_obs_apcp.cfg".  At runtime, Ensemble-Stat (like
all the other MET tools with config files) actually reads multiple
config files:
 (1) Constants file: met-8.0/share/met/ConfigConstants
 (2) Default for the tool: met-
8.0/share/met/EnsembleStatConfig_default
 (3) User-specified one from the command
line:ensemble_stat_model_obs_apcp.cfg

These reads are cumulative.  So the settings you put in your user
config file override the defaults.  Anything you do not override
remains at it's default setting.  And that's what's happening here.
Your config file is missing the ensemble (ens) section which lists the
fields for which you want ensemble means and relative frequencies
computed.  So it's using the default setting which has 3-hourly
accumulated precip (met-8.0/share/met/EnsembleStatConfig_default).

Frankly an empty list probably is a better default value... but it's a
balance between good defaults and good examples.

Please try adding the ens section into your config file and setting it
to an empty list:

*ens = {   ens_thresh = 1.0;   vld_thresh = 1.0;   field = [ ];}*

Hope that helps.  Note that I did not go through your config file
carefully to see if other items are missing.  So it'd be a good idea
to compare yours with the default to think about your config file
choices.

Hope that helps.

Thanks,
John Halley Gotway

On Mon, Jan 28, 2019 at 2:38 PM SARASAMMA, SUSEELA T CTR USAF AFMC
AFLCMC/HBAW-OL via RT <met_help at ucar.edu> wrote:

>
> Mon Jan 28 14:38:15 2019: Request 88674 was acted upon.
> Transaction: Ticket created by suseela.sarasamma.ctr at us.af.mil
>        Queue: met_help
>      Subject: Question on ensemble_stat when using point_obs
>        Owner: Nobody
>   Requestors: suseela.sarasamma.ctr at us.af.mil
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=88674
> >
>
>
> Hi,
>
> I am trying to run ensemble_stat on point observations for Total
> Precipitation (APCP).
>
>
>
> Even though the met config file I am using does not have any A03
level
> in it, the met_log shows it is looking for
>
> APCP/A03 in the GEPS ensemble GFS forecast file. Then I get the
> following WARNING message:
>
>
>
> WARNING: MetGrib2DataFile::data_plane() - No matching record found
for
> 'APCP/A03'
>
> WARNING:
>
> WARNING:
>
> WARNING: process_n_vld() -> ensemble field "APCP/A03" not found in
> file
>
>
"/home/sarasammas/geps_verify/reduced_gribs/2018120500/GFS/GFS_E00_201
> 812050
> 00000_f012.gr2"
>
>
>
> For the run as shown in the attached log file, I was using a
> non-standard grib_table, thinking the observation file was created
> using that grib_table.
>
>
>
> I had no trouble running point_stat on the same pair of forecast
file
> and observation file.
>
>
>
> For your reference I am attaching the log file, my config file and
the
> observation file.
>
>
>
> Thanks
>
> Suseela
>
>
>
>
>

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