[Met_help] [rt.rap.ucar.edu #63092] History for stat_analysis for wind speed threshold using -line_type MPR
John Halley Gotway via RT
met_help at ucar.edu
Mon Oct 14 14:22:56 MDT 2013
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi John,
I'm trying to run stat_analysis to verify wind speed from the outputs
of point_stat but it doesn't seem to be using the threshold option.
I've tried so many different things but these last 5 tries should be
enough to explain what's happening:
What I want to do is to output statistics for wind speeds only above a
certain threshold. It happens that among all the MPR lines written
with point_stat, I have observations with 0.000 values and I don't
trust those values. Therefore I don't want them to be computed, but
it's not what stat_analysis is doing.
For testing purposes I'm running stat_analysis with only 2 point_stat
outputs, which are attached here (only the file of 24th at 12Z has the
zero value obs).
I've tried running stat_analysis with the following:
1) Output CTS using -out_fcst_thresh & -out_obs_thresh >1.0:
stat_analysis -lookin point_stat/*.stat -out
wind_stat_analys/stat_analys_MPR2CTS_tr1 -log stat_analysis.log -v 5
-model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
-line_type MPR -job aggregate_stat -out_line_type CTS -out_fcst_thresh
">1.0" -out_obs_thresh ">1.0" -dump_row "dump_row.out_tr1"
2) Output CTS using -out_fcst_thresh & -out_obs_thresh >0.0
stat_analysis -lookin point_stat/*.stat -out
wind_stat_analys/stat_analys_MPR2CTS_tr0 -log stat_analysis.log -v 5
-model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
-line_type MPR -job aggregate_stat -out_line_type CTS -out_fcst_thresh
">0.0" -out_obs_thresh ">0.0" -dump_row "dump_row.out_tr0"
3) Output CTC using -out_fcst_thresh & -out_obs_thresh >1.0:
stat_analysis -lookin point_stat/*.stat -out
wind_stat_analys/stat_analys_MPR2CTC_tr1 -log stat_analysis.log -v 5
-model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
-line_type MPR -job aggregate_stat -out_line_type CTC -out_fcst_thresh
">1.0" -out_obs_thresh ">1.0" -dump_row "dump_row.out_tr1_ctc"
4) Output CTC using -out_fcst_thresh & -out_obs_thresh >0.0:
stat_analysis -lookin point_stat/*.stat -out
wind_stat_analys/stat_analys_MPR2CTC_tr0 -log stat_analysis.log -v 5
-model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
-line_type MPR -job aggregate_stat -out_line_type CTC -out_fcst_thresh
">0.0" -out_obs_thresh ">0.0" -dump_row "dump_row.out_tr0_ctc"
5) Output CTC using -out_fcst_thresh & -out_obs_thresh >0.0 AND using
-obs_thresh >1.0 -fcst_thresh >1.0
stat_analysis -lookin point_stat/*.stat -out
wind_stat_analys/stat_analys_MPR2CTC_tr1_2 -log stat_analysis.log -v 5
-model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
-line_type MPR -job aggregate_stat -out_line_type CTC -out_fcst_thresh
">1.0" -out_obs_thresh ">1.0" -dump_row "dump_row.out_tr1_ctc_2"
-obs_thresh ">1.0" -fcst_thresh ">1.0"
Both my CTS's outputs show the same values for the calculated
statistics except for the bootstrap lower confidence limits (not
included in this email). The CTC's, both include all the MPR lines
(69) available in the point_stat files for the counts. Both show:
COL_NAME: TOTAL FY_OY FY_ON FN_OY FN_ON
CTC: 69 61 8 0 0
And finally, when I try #5, I get zero STAT lines. See the log lines below:
DEBUG 1: Creating STAT-Analysis output file
"wind_stat_analys/stat_analys_MPR2CTC_tr1_2"
DEBUG 4: Amending default job with command line options: "-model WRF
-fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10 -line_type MPR
-job aggregate_stat -out_line_type CTC -out_fcst_thresh >1.0
-out_obs_thresh >1.0 -dump_row dump_row.out_tr1_ctc_2 -obs_thresh >1.0
-fcst_thresh >1.0"
DEBUG 3: Processing STAT file
"point_stat/point_stat_420000L_20080724_120000V.stat" ... 1 of 2
DEBUG 3: Processing STAT file
"point_stat/point_stat_480000L_20080724_180000V.stat" ... 2 of 2
DEBUG 2: STAT Lines read = 73
DEBUG 2: STAT Lines retained = 0
DEBUG 4:
DEBUG 4: Initializing Job 1 to default job: "-job aggregate_stat
-model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
-fcst_thresh >1.000 -obs_thresh >1.000 -line_type MPR -dump_row
dump_row.out_tr1_ctc_2 -out_line_type CTC -out_fcst_thresh >1.000
-out_obs_thresh >1.000 "
DEBUG 4:
DEBUG 4: Amending Job 1 with options: "-model WRF -fcst_var WIND
-obs_var WIND -fcst_lev Z10 -obs_lev Z10 -line_type MPR -job
aggregate_stat -out_line_type CTC -out_fcst_thresh >1.0
-out_obs_thresh >1.0 -dump_row dump_row.out_tr1_ctc_2 -obs_thresh >1.0
-fcst_thresh >1.0"
DEBUG 4:
DEBUG 4: Amending Job 1 with command line options: "-model WRF
-fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10 -line_type MPR
-job aggregate_stat -out_line_type CTC -out_fcst_thresh >1.0
-out_obs_thresh >1.0 -dump_row dump_row.out_tr1_ctc_2 -obs_thresh >1.0
-fcst_thresh >1.0"
DEBUG 2:
DEBUG 2: Processing Job 1: -job aggregate_stat -model WRF -fcst_var
WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10 -fcst_thresh >1.000
-obs_thresh >1.000 -line_type MPR -dump_row dump_row.out_tr1_ctc_2
-out_line_type CTC -out_fcst_thresh >1.000 -out_obs_thresh >1.000
DEBUG 1: Creating STAT output file "dump_row.out_tr1_ctc_2"
WARNING:
WARNING: do_job_aggr_stat() -> no matching STAT lines found for job:
-job aggregate_stat -model WRF -fcst_var WIND -obs_var WIND -fcst_lev
Z10 -obs_lev Z10 -fcst_thresh >1.000 -obs_thresh >1.000 -line_type MPR
-dump_row dump_row.out_tr1_ctc_2 -out_line_type CTC -out_fcst_thresh
>1.000 -out_obs_thresh >1.000
WARNING:
DEBUG 2: Job 1 used 0 out of 0 STAT lines.
I have also tried changing the configure file for point_stat in order
to remove the 0 value observations ( I tried changing: wind_thresh =
[ >0.1 ]), but they show in the file anyway.
I'm using the MET tools from yellowstone, loaded with module load METv4.0/4.0
Could it be a bug in the code? What can I do to ignore the pairs where
obs-wind=0 ?
Thanks a lot
Maria Eugenia B. Frediani
-------------------------------------------------------------------------------------
PhD Candidate
University of Connecticut
School of Engineering
261 Glenbrook Rd
Storrs, CT 06269
maria.frediani at uconn.edu
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63092] stat_analysis for wind speed threshold using -line_type MPR
From: John Halley Gotway
Time: Thu Sep 19 17:02:58 2013
Maria,
I understand what you're trying to do. I think this just bears a
little more explanation. Let's talk about the first job you list:
-line_type MPR -job aggregate_stat -out_line_type CTS
-out_fcst_thresh ">1.0" -out_obs_thresh ">1.0"
This is telling STAT-Analysis to read in all of the MPR lines from the
STAT files, recompute a 2x2 contingency table using the >1.0
thresholds you've provided, and write out the corresponding
contingency table statistics. I believe it's doing exactly what
you're telling it to do using all 69 MPR lines.
The problem is that you actually want it to do something slightly
different than this. You want it to read in all the MPR lines, throw
out any of the MPR lines where either the forecast or
observation value is <=1.0, and then use the remaining line to compute
a 2x2 contingency table. This can be done using the "-column_min"
option. This is listed in METv4.1/data/config/README.
Try running the following job:
stat_analysis -lookin point_stat/*.stat \
-out wind_stat_analys/stat_analys_MPR2CTS_tr1 -log stat_analysis.log \
-v 5 -model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev
Z10 \
-line_type MPR -job aggregate_stat -out_line_type CTS \
-column_min FCST 1.0 -column_min OBS 1.0 \
-out_fcst_thresh ">1.0" -out_obs_thresh ">1.0" \
-dump_row "dump_row.out_tr1"
The column_min option tells it to filter in input data by applying the
threshold you specify to the column listed. This job only uses 61 of
69 MPR lines because 8 of them have 0's in the OBS column
of the MPR lines.
Hope that helps clarify.
Thanks,
John
On 09/19/2013 02:15 PM, Maria Eugenia via RT wrote:
>
> Thu Sep 19 14:15:21 2013: Request 63092 was acted upon.
> Transaction: Ticket created by maria.frediani at gmail.com
> Queue: met_help
> Subject: stat_analysis for wind speed threshold using
-line_type MPR
> Owner: Nobody
> Requestors: maria.frediani at gmail.com
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63092 >
>
>
> Hi John,
>
> I'm trying to run stat_analysis to verify wind speed from the
outputs
> of point_stat but it doesn't seem to be using the threshold option.
> I've tried so many different things but these last 5 tries should be
> enough to explain what's happening:
>
> What I want to do is to output statistics for wind speeds only above
a
> certain threshold. It happens that among all the MPR lines written
> with point_stat, I have observations with 0.000 values and I don't
> trust those values. Therefore I don't want them to be computed, but
> it's not what stat_analysis is doing.
>
> For testing purposes I'm running stat_analysis with only 2
point_stat
> outputs, which are attached here (only the file of 24th at 12Z has
the
> zero value obs).
>
> I've tried running stat_analysis with the following:
>
> 1) Output CTS using -out_fcst_thresh & -out_obs_thresh >1.0:
> stat_analysis -lookin point_stat/*.stat -out
> wind_stat_analys/stat_analys_MPR2CTS_tr1 -log stat_analysis.log -v 5
> -model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
> -line_type MPR -job aggregate_stat -out_line_type CTS
-out_fcst_thresh
> ">1.0" -out_obs_thresh ">1.0" -dump_row "dump_row.out_tr1"
>
> 2) Output CTS using -out_fcst_thresh & -out_obs_thresh >0.0
> stat_analysis -lookin point_stat/*.stat -out
> wind_stat_analys/stat_analys_MPR2CTS_tr0 -log stat_analysis.log -v 5
> -model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
> -line_type MPR -job aggregate_stat -out_line_type CTS
-out_fcst_thresh
> ">0.0" -out_obs_thresh ">0.0" -dump_row "dump_row.out_tr0"
>
> 3) Output CTC using -out_fcst_thresh & -out_obs_thresh >1.0:
> stat_analysis -lookin point_stat/*.stat -out
> wind_stat_analys/stat_analys_MPR2CTC_tr1 -log stat_analysis.log -v 5
> -model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
> -line_type MPR -job aggregate_stat -out_line_type CTC
-out_fcst_thresh
> ">1.0" -out_obs_thresh ">1.0" -dump_row "dump_row.out_tr1_ctc"
>
> 4) Output CTC using -out_fcst_thresh & -out_obs_thresh >0.0:
> stat_analysis -lookin point_stat/*.stat -out
> wind_stat_analys/stat_analys_MPR2CTC_tr0 -log stat_analysis.log -v 5
> -model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
> -line_type MPR -job aggregate_stat -out_line_type CTC
-out_fcst_thresh
> ">0.0" -out_obs_thresh ">0.0" -dump_row "dump_row.out_tr0_ctc"
>
> 5) Output CTC using -out_fcst_thresh & -out_obs_thresh >0.0 AND
using
> -obs_thresh >1.0 -fcst_thresh >1.0
> stat_analysis -lookin point_stat/*.stat -out
> wind_stat_analys/stat_analys_MPR2CTC_tr1_2 -log stat_analysis.log -v
5
> -model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
> -line_type MPR -job aggregate_stat -out_line_type CTC
-out_fcst_thresh
> ">1.0" -out_obs_thresh ">1.0" -dump_row "dump_row.out_tr1_ctc_2"
> -obs_thresh ">1.0" -fcst_thresh ">1.0"
>
> Both my CTS's outputs show the same values for the calculated
> statistics except for the bootstrap lower confidence limits (not
> included in this email). The CTC's, both include all the MPR lines
> (69) available in the point_stat files for the counts. Both show:
>
> COL_NAME: TOTAL FY_OY FY_ON FN_OY FN_ON
> CTC: 69 61 8 0 0
>
>
> And finally, when I try #5, I get zero STAT lines. See the log lines
below:
>
> DEBUG 1: Creating STAT-Analysis output file
> "wind_stat_analys/stat_analys_MPR2CTC_tr1_2"
> DEBUG 4: Amending default job with command line options: "-model WRF
> -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10 -line_type
MPR
> -job aggregate_stat -out_line_type CTC -out_fcst_thresh >1.0
> -out_obs_thresh >1.0 -dump_row dump_row.out_tr1_ctc_2 -obs_thresh
>1.0
> -fcst_thresh >1.0"
> DEBUG 3: Processing STAT file
> "point_stat/point_stat_420000L_20080724_120000V.stat" ... 1 of 2
>
> DEBUG 3: Processing STAT file
> "point_stat/point_stat_480000L_20080724_180000V.stat" ... 2 of 2
>
> DEBUG 2: STAT Lines read = 73
> DEBUG 2: STAT Lines retained = 0
> DEBUG 4:
> DEBUG 4: Initializing Job 1 to default job: "-job aggregate_stat
> -model WRF -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10
> -fcst_thresh >1.000 -obs_thresh >1.000 -line_type MPR -dump_row
> dump_row.out_tr1_ctc_2 -out_line_type CTC -out_fcst_thresh >1.000
> -out_obs_thresh >1.000 "
> DEBUG 4:
> DEBUG 4: Amending Job 1 with options: "-model WRF -fcst_var WIND
> -obs_var WIND -fcst_lev Z10 -obs_lev Z10 -line_type MPR -job
> aggregate_stat -out_line_type CTC -out_fcst_thresh >1.0
> -out_obs_thresh >1.0 -dump_row dump_row.out_tr1_ctc_2 -obs_thresh
>1.0
> -fcst_thresh >1.0"
> DEBUG 4:
> DEBUG 4: Amending Job 1 with command line options: "-model WRF
> -fcst_var WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10 -line_type
MPR
> -job aggregate_stat -out_line_type CTC -out_fcst_thresh >1.0
> -out_obs_thresh >1.0 -dump_row dump_row.out_tr1_ctc_2 -obs_thresh
>1.0
> -fcst_thresh >1.0"
> DEBUG 2:
> DEBUG 2: Processing Job 1: -job aggregate_stat -model WRF -fcst_var
> WIND -obs_var WIND -fcst_lev Z10 -obs_lev Z10 -fcst_thresh >1.000
> -obs_thresh >1.000 -line_type MPR -dump_row dump_row.out_tr1_ctc_2
> -out_line_type CTC -out_fcst_thresh >1.000 -out_obs_thresh >1.000
> DEBUG 1: Creating STAT output file "dump_row.out_tr1_ctc_2"
> WARNING:
> WARNING: do_job_aggr_stat() -> no matching STAT lines found for job:
> -job aggregate_stat -model WRF -fcst_var WIND -obs_var WIND
-fcst_lev
> Z10 -obs_lev Z10 -fcst_thresh >1.000 -obs_thresh >1.000 -line_type
MPR
> -dump_row dump_row.out_tr1_ctc_2 -out_line_type CTC -out_fcst_thresh
>> 1.000 -out_obs_thresh >1.000
> WARNING:
> DEBUG 2: Job 1 used 0 out of 0 STAT lines.
>
>
> I have also tried changing the configure file for point_stat in
order
> to remove the 0 value observations ( I tried changing: wind_thresh
=
> [ >0.1 ]), but they show in the file anyway.
>
> I'm using the MET tools from yellowstone, loaded with module load
METv4.0/4.0
>
> Could it be a bug in the code? What can I do to ignore the pairs
where
> obs-wind=0 ?
>
> Thanks a lot
>
> Maria Eugenia B. Frediani
>
-------------------------------------------------------------------------------------
> PhD Candidate
> University of Connecticut
> School of Engineering
> 261 Glenbrook Rd
> Storrs, CT 06269
> maria.frediani at uconn.edu
>
------------------------------------------------
More information about the Met_help
mailing list