[Met_help] [rt.rap.ucar.edu #88670] History for Stat_analysis Use of Climo
John Halley Gotway via RT
met_help at ucar.edu
Tue Jul 9 12:07:07 MDT 2019
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
John, with your previous help I have progressed to running stat analysis on
my HIRA data. And I am running into another oddity. When I generate PSTD
files using Point Stat on HIRA data, the PSTD files show climo data in the
biercl field and the BSS is not NA. However, if I run stat_analysis on the
STAT files with linetpe MPR, and out line type PSTD, the briercl is NA and
BSS is NA. The BSS Simple however is the same as what I got in the PSTD
file running point stat. I sent the STAT, PCT, and PSTD files to you view
AMR Safe. Attached Is the configuration file I am using. Why don't I get
a valid briercl and BSS when running the data through stat_analysis?
Thanks
Bob
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Stat_analysis Use of Climo
From: John Halley Gotway
Time: Mon Jan 28 12:09:30 2019
Bob,
First, I'm not receiving anything through ARM SAFE... both you and
Matt
Sittel have been trying to send me data that way, but I haven't seen
anything. I'm not sure what's going on there. But I do have some
sample
data that I used to replicate this behavior.
As you're finding, we haven't run this combination of HiRA and
Stat-Analysis in the past. So we're running into all sorts of issues
that
haven't arisen previously!
It took me a while to piece together the logic but I think this is
what's
happening...
(1) Going from met-6.0 to met-6.1, we changed the columns of the MPR
line
type by adding more climo info:
met-6.0: *FCST OBS CLIMO OBS_QC*
met-6.1: *FCST OBS OBS_QC CLIMO_MEAN CLIMO_STDEV CLIMO_CDF*
(2) In doing so, we introduced a bug in met-6.1 by failing to update
the
function which reads these MPR lines back in (parse_mpr_line()):
* m_data.climo = atof(l.get_item("CLIMO"));*
It's continuing to look for a column named *CLIMO* instead of columns
for *CLIMO_MEAN,
CLIMO_STDEV, *and *CLIMO_CDF*.
(3) We fixed this bug in met-8.0. So when I run met-8.0 stat_analysis
to
read MPR and write PSTD, I do get a climatology brier score (BRIERCL)
and a
skill score (BSS). But using met-6.1, I do not because Stat-Analysis
isn't
reading the climo values.
So what do we do?
- We could fix the met-6.1 code as follows. On line 297 of the file
src/tools/core/stat_analysis/parse_stat_line.cc, change "CLIMO" to
"CLIMO_MEAN":
* 297 m_data.climo = atof(l.get_item("CLIMO_MEAN"));*
- Or we could hack the data by renaming the "CLIMO_MEAN" column to
"CLIMO".
# Run Stat-Analysis to read your .stat file and filter the MPR
lines to
an output file
* stat_analysis -lookin input.stat -job filter -line_type MPR
-interp_mthd NBRHD -dump_row temp_mpr.txt*
# Modify the CLIMO_MEAN column name
*cat* *temp_mpr.txt | sed 's/CLIMO_MEAN/CLIMO/g' > patch_mpr.txt*
# Run the Stat-Analysis job on the result
* stat_analysis -lookin patch_mpr.txt -config
STATAnalysisConfig_hira*
Sorry for all this trouble. These pesky little bugs can have long-
lasting
ramifications.
Thanks,
John
On Mon, Jan 28, 2019 at 10:35 AM robert.craig.2 at us.af.mil via RT <
met_help at ucar.edu> wrote:
>
> Mon Jan 28 10:35:20 2019: Request 88670 was acted upon.
> Transaction: Ticket created by robert.craig.2 at us.af.mil
> Queue: met_help
> Subject: Stat_analysis Use of Climo
> Owner: Nobody
> Requestors: robert.craig.2 at us.af.mil
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=88670 >
>
>
> John, with your previous help I have progressed to running stat
analysis on
> my HIRA data. And I am running into another oddity. When I
generate PSTD
> files using Point Stat on HIRA data, the PSTD files show climo data
in the
> biercl field and the BSS is not NA. However, if I run stat_analysis
on the
> STAT files with linetpe MPR, and out line type PSTD, the briercl is
NA and
> BSS is NA. The BSS Simple however is the same as what I got in the
PSTD
> file running point stat. I sent the STAT, PCT, and PSTD files to
you view
> AMR Safe. Attached Is the configuration file I am using. Why
don't I get
> a valid briercl and BSS when running the data through stat_analysis?
>
>
>
> Thanks
>
> Bob
>
>
>
------------------------------------------------
Subject: RE: [Non-DoD Source] Re: [rt.rap.ucar.edu #88670] Stat_analysis Use of Climo
From: robert.craig.2 at us.af.mil
Time: Mon Jan 28 12:16:29 2019
Thanks for the explanation. I let Matt know that ARL Safe is not
working any more. I will try the hack but am looking forward to MET
8.0.
Thanks
Bob
-----Original Message-----
From: John Halley Gotway via RT <met_help at ucar.edu>
Sent: Monday, January 28, 2019 1:10 PM
To: CRAIG, ROBERT J GS-12 USAF ACC 16 WS/WXN
<robert.craig.2 at us.af.mil>
Cc: SITTEL, MATTHEW C CTR USAF AFMC AFLCMC/HBAW-OL
<matthew.sittel.ctr at us.af.mil>
Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #88670] Stat_analysis
Use of Climo
Bob,
First, I'm not receiving anything through ARM SAFE... both you and
Matt Sittel have been trying to send me data that way, but I haven't
seen anything. I'm not sure what's going on there. But I do have
some sample data that I used to replicate this behavior.
As you're finding, we haven't run this combination of HiRA and Stat-
Analysis in the past. So we're running into all sorts of issues that
haven't arisen previously!
It took me a while to piece together the logic but I think this is
what's happening...
(1) Going from met-6.0 to met-6.1, we changed the columns of the MPR
line type by adding more climo info:
met-6.0: *FCST OBS CLIMO OBS_QC*
met-6.1: *FCST OBS OBS_QC CLIMO_MEAN CLIMO_STDEV CLIMO_CDF*
(2) In doing so, we introduced a bug in met-6.1 by failing to update
the function which reads these MPR lines back in (parse_mpr_line()):
* m_data.climo = atof(l.get_item("CLIMO"));*
It's continuing to look for a column named *CLIMO* instead of columns
for *CLIMO_MEAN, CLIMO_STDEV, *and *CLIMO_CDF*.
(3) We fixed this bug in met-8.0. So when I run met-8.0 stat_analysis
to read MPR and write PSTD, I do get a climatology brier score
(BRIERCL) and a skill score (BSS). But using met-6.1, I do not
because Stat-Analysis isn't reading the climo values.
So what do we do?
- We could fix the met-6.1 code as follows. On line 297 of the file
src/tools/core/stat_analysis/parse_stat_line.cc, change "CLIMO" to
"CLIMO_MEAN":
* 297 m_data.climo = atof(l.get_item("CLIMO_MEAN"));*
- Or we could hack the data by renaming the "CLIMO_MEAN" column to
"CLIMO".
# Run Stat-Analysis to read your .stat file and filter the MPR
lines to an output file
* stat_analysis -lookin input.stat -job filter -line_type MPR
-interp_mthd NBRHD -dump_row temp_mpr.txt*
# Modify the CLIMO_MEAN column name
*cat* *temp_mpr.txt | sed 's/CLIMO_MEAN/CLIMO/g' > patch_mpr.txt*
# Run the Stat-Analysis job on the result
* stat_analysis -lookin patch_mpr.txt -config
STATAnalysisConfig_hira*
Sorry for all this trouble. These pesky little bugs can have long-
lasting ramifications.
Thanks,
John
On Mon, Jan 28, 2019 at 10:35 AM robert.craig.2 at us.af.mil via RT <
met_help at ucar.edu> wrote:
>
> Mon Jan 28 10:35:20 2019: Request 88670 was acted upon.
> Transaction: Ticket created by robert.craig.2 at us.af.mil
> Queue: met_help
> Subject: Stat_analysis Use of Climo
> Owner: Nobody
> Requestors: robert.craig.2 at us.af.mil
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=88670
> >
>
>
> John, with your previous help I have progressed to running stat
> analysis on my HIRA data. And I am running into another oddity.
When
> I generate PSTD files using Point Stat on HIRA data, the PSTD files
> show climo data in the biercl field and the BSS is not NA. However,
> if I run stat_analysis on the STAT files with linetpe MPR, and out
> line type PSTD, the briercl is NA and BSS is NA. The BSS Simple
however is the same as what I got in the PSTD
> file running point stat. I sent the STAT, PCT, and PSTD files to
you view
> AMR Safe. Attached Is the configuration file I am using. Why
don't I get
> a valid briercl and BSS when running the data through stat_analysis?
>
>
>
> Thanks
>
> Bob
>
>
>
------------------------------------------------
More information about the Met_help
mailing list