[Met_help] [rt.rap.ucar.edu #57995] History for Point stat Clarification.
Paul Oldenburg via RT
met_help at ucar.edu
Wed Sep 26 12:36:18 MDT 2012
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi Paul, I am doing f/c verf of Rainfall for 122 days using MET. I run WRF for 48hrs FC. 1. I am doing verf of RF using point observations. I have run point stat tool and now I am trying with stat-analysis tool. 2. I have to generate the products like ME, MAE, RMSe etc for my domain for 24 and 48hrs. 3. for this, I have kept 21 stat files in bash-3.2$ cd statfiles/
/oprn/model/wrf3/utils/met/METv3.0/geeta_test/statfiles
bash-3.2$ ls -l
total 3288
-rwxr-xr-x 1 model staff 91922 Jun 11 14:09 point_stat_gt0.0_270000L_20110612_030000V.stat
-rwxr-xr-x 1 model staff 91007 Jun 11 14:09 point_stat_gt0.0_270000L_20110613_030000V.stat
-rwxr-xr-x 1 model staff 91912 Jun 11 14:09 point_stat_gt0.0_270000L_20110615_030000V.stat
-rwxr-xr-x 1 model staff 92219 Jun 11 14:09 point_stat_gt0.0_270000L_20110616_030000V.stat
-rwxr-xr-x 1 model staff 92834 Jun 11 14:09 point_stat_gt0.0_270000L_20110618_030000V.stat
-rwxr-xr-x 1 model staff 90426 Jun 11 14:09 point_stat_gt0.0_270000L_20110621_030000V.stat
-rwxr-xr-x 1 model staff 92539 Jun 11 14:09 point_stat_gt0.0_270000L_20110713_030000V.stat
-rwxr-xr-x 1 model staff 93487 Jun 11 14:09 point_stat_gt0.0_270000L_20110714_030000V.stat
-rwxr-xr-x 1 model staff 214 Jun 11 14:09 point_stat_gt0.0_270000L_20110715_030000V.stat
-rwxr-xr-x 1 model staff 92852 Jun 11 14:09 point_stat_gt0.0_270000L_20110716_030000V.stat
-rwxr-xr-x 1 model staff 92855 Jun 11 14:09 point_stat_gt0.0_270000L_20110719_030000V.stat
-rwxr-xr-x 1 model staff 90045 Jun 11 14:09 point_stat_gt0.0_270000L_20110720_030000V.stat
-rwxr-xr-x 1 model staff 91585 Jun 11 14:09 point_stat_gt0.0_270000L_20110721_030000V.stat
-rwxr-xr-x 1 model staff 89775 Jun 11 14:09 point_stat_gt0.0_270000L_20110723_030000V.stat
-rwxr-xr-x 1 model staff 89765 Jun 11 14:09 point_stat_gt0.0_270000L_20110724_030000V.stat
-rwxr-xr-x 1 model staff 86769 Jun 11 14:09 point_stat_gt0.0_270000L_20110728_030000V.stat
-rwxr-xr-x 1 model staff 88646 Jun 11 14:09 point_stat_gt0.0_270000L_20110729_030000V.stat
-rwxr-xr-x 1 model staff 90156 Jun 11 14:09 point_stat_gt0.0_270000L_20110730_030000V.stat
-rwxr-xr-x 1 model staff 89462 Jun 11 14:09 point_stat_gt0.0_270000L_20110731_030000V.stat
drwxr-xr-x 2 model staff 8192 Jun 14 11:46 tempo
bash-3.2$ ls -l | wc -l
21
4. I have run stat-tool using the command/oprn/model/wrf3/utils/met/METv3.0/bin/stat_analysis \
-job summary -line_type CNT -column RMSE \
-dump_row summary_test.stat \
-lookin statfiles \
-fcst_var APCP_24 -fcst_lev A24 -v 2 5. It runs fine but I have following query. In the O/P file, sumary_test_type, I have 214 lines containing CNT. But I have 21 files as input. Also, for example 20110612_03 (one of the dates) is reapeated in the o/p file 6-7 times. Also the data corresponsing to this day is also different. I thought that stat-tool extracts all the RMSE columns from the line CNT and puts them at one place. Pls clarify on this how am I getting so many lines. I am attaching the O/p file. 6. I had tried with aggregate tool, but that does not seem to work very efficiently. Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110719_030000V.stat" .... 106 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110720_030000V.stat" .... 107 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110721_030000V.stat" .... 108 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110723_030000V.stat" .... 109 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110724_030000V.stat" .... 110 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110728_030000V.stat" .... 111 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110729_030000V.stat" .... 112 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110730_030000V.stat" .... 113 of 114
Processing STAT file "statfiles/tempo/point_stat_ge7.6_270000L_20110731_030000V.stat" .... 114 of 114
STAT Lines read = 62698
STAT Lines retained = 62698Processing Job 1: -job aggregate_stat -line_type MPR -dump_row agg_sta.dat -out_line_type CNT -out_alpha 0.050000 -boot_interval 1 -boot_rep_prop 1.000000 -n_boot_rep 1000 -boot_rng mt19937 -boot_seed -rank_corr_flag 1
Creating STAT output file "agg_sta.dat"
DataLine::get_item(int) -> range check errorbash-3.2$ I have the aggregate file but it jus shows the MPR lines for each date????
How to proceed further????
geeta
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #57995] Point stat Clarification.
From: Paul Oldenburg
Time: Fri Aug 24 09:08:51 2012
Geeta,
Your point_stat files may contain more than one CNT line per file.
That depends on how you configured point_stat. For
example, verifying multiple forecast variables, thresholds, masking
regions or interpolation types will all cause
point_stat to write more than one CNT line to an output file. This
would also explain the appearance of redundant dates
in the stat_analysis output.
As for your questions about aggregating the MPR lines into a CNT line,
the slow execution is caused by the bootstrapping
process. This is an unavoidable drawback to the bootstrapping
process. If you want to increase performance, and you
are not interested in the bootstrap confidence intervals, you can
disable this feature by omitting the boot parameters
from your stat_analysis command. Although it is difficult to read in
your email, the error that you reported may be
caused by the fact that the -boot_seed parameter requires an argument
and you do not appear to have one specified.
I hope this answers your questions. If you have any further questions
on this matter, it might be easiest for me to
help you if you bundle your point_stat output files into a tarball and
upload it to our FTP site using the following
instructions:
http://www.dtcenter.org/met/users/support/met_help.php#ftp
Thanks,
Paul
On 08/24/2012 02:58 AM, Geeta Geeta via RT wrote:
>
> Fri Aug 24 02:58:57 2012: Request 57995 was acted upon.
> Transaction: Ticket created by geeta124 at hotmail.com
> Queue: met_help
> Subject: Point stat Clarification.
> Owner: Nobody
> Requestors: geeta124 at hotmail.com
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=57995 >
>
>
>
> Hi Paul, I am doing f/c verf of Rainfall for 122 days using MET. I
run WRF for 48hrs FC. 1. I am doing verf of RF using point
observations. I have run point stat tool and now I am trying with
stat-analysis tool. 2. I have to generate the products like ME, MAE,
RMSe etc for my domain for 24 and 48hrs. 3. for this, I have kept 21
stat files in bash-3.2$ cd statfiles/
> /oprn/model/wrf3/utils/met/METv3.0/geeta_test/statfiles
> bash-3.2$ ls -l
> total 3288
> -rwxr-xr-x 1 model staff 91922 Jun 11 14:09
point_stat_gt0.0_270000L_20110612_030000V.stat
> -rwxr-xr-x 1 model staff 91007 Jun 11 14:09
point_stat_gt0.0_270000L_20110613_030000V.stat
> -rwxr-xr-x 1 model staff 91912 Jun 11 14:09
point_stat_gt0.0_270000L_20110615_030000V.stat
> -rwxr-xr-x 1 model staff 92219 Jun 11 14:09
point_stat_gt0.0_270000L_20110616_030000V.stat
> -rwxr-xr-x 1 model staff 92834 Jun 11 14:09
point_stat_gt0.0_270000L_20110618_030000V.stat
> -rwxr-xr-x 1 model staff 90426 Jun 11 14:09
point_stat_gt0.0_270000L_20110621_030000V.stat
> -rwxr-xr-x 1 model staff 92539 Jun 11 14:09
point_stat_gt0.0_270000L_20110713_030000V.stat
> -rwxr-xr-x 1 model staff 93487 Jun 11 14:09
point_stat_gt0.0_270000L_20110714_030000V.stat
> -rwxr-xr-x 1 model staff 214 Jun 11 14:09
point_stat_gt0.0_270000L_20110715_030000V.stat
> -rwxr-xr-x 1 model staff 92852 Jun 11 14:09
point_stat_gt0.0_270000L_20110716_030000V.stat
> -rwxr-xr-x 1 model staff 92855 Jun 11 14:09
point_stat_gt0.0_270000L_20110719_030000V.stat
> -rwxr-xr-x 1 model staff 90045 Jun 11 14:09
point_stat_gt0.0_270000L_20110720_030000V.stat
> -rwxr-xr-x 1 model staff 91585 Jun 11 14:09
point_stat_gt0.0_270000L_20110721_030000V.stat
> -rwxr-xr-x 1 model staff 89775 Jun 11 14:09
point_stat_gt0.0_270000L_20110723_030000V.stat
> -rwxr-xr-x 1 model staff 89765 Jun 11 14:09
point_stat_gt0.0_270000L_20110724_030000V.stat
> -rwxr-xr-x 1 model staff 86769 Jun 11 14:09
point_stat_gt0.0_270000L_20110728_030000V.stat
> -rwxr-xr-x 1 model staff 88646 Jun 11 14:09
point_stat_gt0.0_270000L_20110729_030000V.stat
> -rwxr-xr-x 1 model staff 90156 Jun 11 14:09
point_stat_gt0.0_270000L_20110730_030000V.stat
> -rwxr-xr-x 1 model staff 89462 Jun 11 14:09
point_stat_gt0.0_270000L_20110731_030000V.stat
> drwxr-xr-x 2 model staff 8192 Jun 14 11:46 tempo
> bash-3.2$ ls -l | wc -l
> 21
>
> 4. I have run stat-tool using the
command/oprn/model/wrf3/utils/met/METv3.0/bin/stat_analysis \
> -job summary -line_type CNT -column RMSE \
> -dump_row summary_test.stat \
> -lookin statfiles \
> -fcst_var APCP_24 -fcst_lev A24 -v 2 5. It runs fine but I have
following query. In the O/P file, sumary_test_type, I have 214 lines
containing CNT. But I have 21 files as input. Also, for example
20110612_03 (one of the dates) is reapeated in the o/p file 6-7 times.
Also the data corresponsing to this day is also different. I thought
that stat-tool extracts all the RMSE columns from the line CNT and
puts them at one place. Pls clarify on this how am I getting so many
lines. I am attaching the O/p file. 6. I had tried with aggregate
tool, but that does not seem to work very efficiently. Processing STAT
file "statfiles/tempo/point_stat_ge7.6_270000L_20110719_030000V.stat"
.... 106 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110720_030000V.stat" ....
107 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110721_030000V.stat" ....
108 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110723_030000V.stat" ....
109 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110724_030000V.stat" ....
110 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110728_030000V.stat" ....
111 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110729_030000V.stat" ....
112 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110730_030000V.stat" ....
113 of 114
> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110731_030000V.stat" ....
114 of 114
> STAT Lines read = 62698
> STAT Lines retained = 62698Processing Job 1: -job aggregate_stat
-line_type MPR -dump_row agg_sta.dat -out_line_type CNT -out_alpha
0.050000 -boot_interval 1 -boot_rep_prop 1.000000 -n_boot_rep 1000
-boot_rng mt19937 -boot_seed -rank_corr_flag 1
> Creating STAT output file "agg_sta.dat"
> DataLine::get_item(int) -> range check errorbash-3.2$ I have the
aggregate file but it jus shows the MPR lines for each date????
> How to proceed further????
>
> geeta
>
------------------------------------------------
Subject: Point stat Clarification.
From: Geeta Geeta
Time: Mon Aug 27 00:28:25 2012
Hi paul,
I am using following commands for the aggregate output.
/oprn/model/wrf3/utils/met/METv3.0/bin/stat_analysis \
-job aggregate_stat -line_type MPR -out_line_type CNT \
-dump_row agg_sta.dat \
-lookin statfiles -v 4Actually, there is not much info reg this in
the PDF file containing the documantation. I am attaching the
aggregate output after I run the above command. How do I obtain Me,
MAE and RMSE etc from these stat files???????????????GeetaFrom:
geeta124 at hotmail.com
To: met_help at ucar.edu
Subject: RE: [rt.rap.ucar.edu #57995] Point stat Clarification.
Date: Mon, 27 Aug 2012 11:38:52 +0530
Hi Paul,
1. I had made a directory called statfiles where I had kept all the
statfiles.
bash-3.2$ ls
agg_sta.dat out sa_test2.sh
summary_test.stat
geeta_test.tar sa_test.sh statfiles
bash-3.2$
there were subdirectories which also had the stat files. So I have
removed the subdirectory and Now I am getting only that number of CNT
entries as the No of files. So this problem looks like fixed.
2. So summary is like collection of the results in one file as
specified by the filter.
3. I wish to know that the job aggregate willl take the mean of all
the days output???? or how does it give me a single figure of RMSE for
100 days of verification of RF.
4. So If I run the summary job and then extract the RMSE column from
the summary o/p file, and then take a mean, will I arrive at the same
result as that of AGGREGATE????? basically, I wish to know, what
exactly the aggregate does????
5. I am sending you the statfiles. Kindly have a look.
Thanks.
geeta
> Subject: Re: [rt.rap.ucar.edu #57995] Point stat Clarification.
> From: met_help at ucar.edu
> To: geeta124 at hotmail.com
> CC: met_help-owner at mailman.ucar.edu
> Date: Fri, 24 Aug 2012 09:08:52 -0600
>
> Geeta,
>
> Your point_stat files may contain more than one CNT line per file.
That depends on how you configured point_stat. For
> example, verifying multiple forecast variables, thresholds, masking
regions or interpolation types will all cause
> point_stat to write more than one CNT line to an output file. This
would also explain the appearance of redundant dates
> in the stat_analysis output.
>
> As for your questions about aggregating the MPR lines into a CNT
line, the slow execution is caused by the bootstrapping
> process. This is an unavoidable drawback to the bootstrapping
process. If you want to increase performance, and you
> are not interested in the bootstrap confidence intervals, you can
disable this feature by omitting the boot parameters
> from your stat_analysis command. Although it is difficult to read
in your email, the error that you reported may be
> caused by the fact that the -boot_seed parameter requires an
argument and you do not appear to have one specified.
>
> I hope this answers your questions. If you have any further
questions on this matter, it might be easiest for me to
> help you if you bundle your point_stat output files into a tarball
and upload it to our FTP site using the following
> instructions:
>
> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>
> Thanks,
>
> Paul
>
>
>
> On 08/24/2012 02:58 AM, Geeta Geeta via RT wrote:
> >
> > Fri Aug 24 02:58:57 2012: Request 57995 was acted upon.
> > Transaction: Ticket created by geeta124 at hotmail.com
> > Queue: met_help
> > Subject: Point stat Clarification.
> > Owner: Nobody
> > Requestors: geeta124 at hotmail.com
> > Status: new
> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=57995 >
> >
> >
> >
> > Hi Paul, I am doing f/c verf of Rainfall for 122 days using MET. I
run WRF for 48hrs FC. 1. I am doing verf of RF using point
observations. I have run point stat tool and now I am trying with
stat-analysis tool. 2. I have to generate the products like ME, MAE,
RMSe etc for my domain for 24 and 48hrs. 3. for this, I have kept 21
stat files in bash-3.2$ cd statfiles/
> > /oprn/model/wrf3/utils/met/METv3.0/geeta_test/statfiles
> > bash-3.2$ ls -l
> > total 3288
> > -rwxr-xr-x 1 model staff 91922 Jun 11 14:09
point_stat_gt0.0_270000L_20110612_030000V.stat
> > -rwxr-xr-x 1 model staff 91007 Jun 11 14:09
point_stat_gt0.0_270000L_20110613_030000V.stat
> > -rwxr-xr-x 1 model staff 91912 Jun 11 14:09
point_stat_gt0.0_270000L_20110615_030000V.stat
> > -rwxr-xr-x 1 model staff 92219 Jun 11 14:09
point_stat_gt0.0_270000L_20110616_030000V.stat
> > -rwxr-xr-x 1 model staff 92834 Jun 11 14:09
point_stat_gt0.0_270000L_20110618_030000V.stat
> > -rwxr-xr-x 1 model staff 90426 Jun 11 14:09
point_stat_gt0.0_270000L_20110621_030000V.stat
> > -rwxr-xr-x 1 model staff 92539 Jun 11 14:09
point_stat_gt0.0_270000L_20110713_030000V.stat
> > -rwxr-xr-x 1 model staff 93487 Jun 11 14:09
point_stat_gt0.0_270000L_20110714_030000V.stat
> > -rwxr-xr-x 1 model staff 214 Jun 11 14:09
point_stat_gt0.0_270000L_20110715_030000V.stat
> > -rwxr-xr-x 1 model staff 92852 Jun 11 14:09
point_stat_gt0.0_270000L_20110716_030000V.stat
> > -rwxr-xr-x 1 model staff 92855 Jun 11 14:09
point_stat_gt0.0_270000L_20110719_030000V.stat
> > -rwxr-xr-x 1 model staff 90045 Jun 11 14:09
point_stat_gt0.0_270000L_20110720_030000V.stat
> > -rwxr-xr-x 1 model staff 91585 Jun 11 14:09
point_stat_gt0.0_270000L_20110721_030000V.stat
> > -rwxr-xr-x 1 model staff 89775 Jun 11 14:09
point_stat_gt0.0_270000L_20110723_030000V.stat
> > -rwxr-xr-x 1 model staff 89765 Jun 11 14:09
point_stat_gt0.0_270000L_20110724_030000V.stat
> > -rwxr-xr-x 1 model staff 86769 Jun 11 14:09
point_stat_gt0.0_270000L_20110728_030000V.stat
> > -rwxr-xr-x 1 model staff 88646 Jun 11 14:09
point_stat_gt0.0_270000L_20110729_030000V.stat
> > -rwxr-xr-x 1 model staff 90156 Jun 11 14:09
point_stat_gt0.0_270000L_20110730_030000V.stat
> > -rwxr-xr-x 1 model staff 89462 Jun 11 14:09
point_stat_gt0.0_270000L_20110731_030000V.stat
> > drwxr-xr-x 2 model staff 8192 Jun 14 11:46 tempo
> > bash-3.2$ ls -l | wc -l
> > 21
> >
> > 4. I have run stat-tool using the
command/oprn/model/wrf3/utils/met/METv3.0/bin/stat_analysis \
> > -job summary -line_type CNT -column RMSE \
> > -dump_row summary_test.stat \
> > -lookin statfiles \
> > -fcst_var APCP_24 -fcst_lev A24 -v 2 5. It runs fine but I have
following query. In the O/P file, sumary_test_type, I have 214 lines
containing CNT. But I have 21 files as input. Also, for example
20110612_03 (one of the dates) is reapeated in the o/p file 6-7 times.
Also the data corresponsing to this day is also different. I thought
that stat-tool extracts all the RMSE columns from the line CNT and
puts them at one place. Pls clarify on this how am I getting so many
lines. I am attaching the O/p file. 6. I had tried with aggregate
tool, but that does not seem to work very efficiently. Processing STAT
file "statfiles/tempo/point_stat_ge7.6_270000L_20110719_030000V.stat"
.... 106 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110720_030000V.stat" ....
107 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110721_030000V.stat" ....
108 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110723_030000V.stat" ....
109 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110724_030000V.stat" ....
110 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110728_030000V.stat" ....
111 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110729_030000V.stat" ....
112 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110730_030000V.stat" ....
113 of 114
> > Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110731_030000V.stat" ....
114 of 114
> > STAT Lines read = 62698
> > STAT Lines retained = 62698Processing Job 1: -job aggregate_stat
-line_type MPR -dump_row agg_sta.dat -out_line_type CNT -out_alpha
0.050000 -boot_interval 1 -boot_rep_prop 1.000000 -n_boot_rep 1000
-boot_rng mt19937 -boot_seed -rank_corr_flag 1
> > Creating STAT output file "agg_sta.dat"
> > DataLine::get_item(int) -> range check errorbash-3.2$ I have
the aggregate file but it jus shows the MPR lines for each date????
> > How to proceed further????
> >
> > geeta
> >
>
>
------------------------------------------------
Subject: Point stat Clarification.
From: Paul Oldenburg
Time: Mon Aug 27 09:03:10 2012
Geeta,
If you are only interested in calculating single RMSE, ME and MAE
values for the matched pair data (MPR lines) that you
generated using point_stat, then the aggregate_stat job is the correct
job to use. The summary job does something else
- it gives distribution information for the values of a single
statistic. Again, I cannot explain the behavior on your
system without testing the point_stat output files that you are using.
I did create a stat_analysis job that aggregates
all of the MPR lines in the agg_sta.dat file that you sent. I
attached the output file showing the RMSE, ME and MAE
that was calculated. The commands that I used are shown below. It is
important to make sure that the only file with
the .stat suffix present in the directory tree where the file is
executed is agg_sta.stat. Please let me know if you
have any other questions.
Paul
$ mv agg_sta.dat agg_sta.stat # change the suffix of
the input file to .stat
$ $MET_BASE/bin/stat_analysis -lookin . -job aggregate_stat -line_type
MPR -out_line_type CNT \
-out agg_sta.out -n_boot_rep 0 -v 4Creating STAT-Analysis output
file "agg_sta.out"
Processing STAT file "./agg_sta.stat" ... 1 of 1
STAT Lines read = 11521
STAT Lines retained = 11521
Processing Job 1: -job aggregate_stat -line_type MPR -out_line_type
CNT -out_alpha 0.050000 -boot_interval 1
-boot_rep_prop 1.000000 -n_boot_rep 0 -boot_rng mt19937 -boot_seed
-rank_corr_flag 1
GSL_RNG_TYPE=mt19937
GSL_RNG_SEED=18446744072460319568
Job 1 used 11521 out of 11521 STAT lines.
On 08/27/2012 12:28 AM, Geeta Geeta via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=57995 >
>
>
> Hi paul,
> I am using following commands for the aggregate output.
> /oprn/model/wrf3/utils/met/METv3.0/bin/stat_analysis \
> -job aggregate_stat -line_type MPR -out_line_type CNT \
> -dump_row agg_sta.dat \
> -lookin statfiles -v 4Actually, there is not much info reg this
in the PDF file containing the documantation. I am attaching the
aggregate output after I run the above command. How do I obtain Me,
MAE and RMSE etc from these stat files???????????????GeetaFrom:
geeta124 at hotmail.com
> To: met_help at ucar.edu
> Subject: RE: [rt.rap.ucar.edu #57995] Point stat Clarification.
> Date: Mon, 27 Aug 2012 11:38:52 +0530
>
>
>
>
>
> Hi Paul,
> 1. I had made a directory called statfiles where I had kept all the
statfiles.
> bash-3.2$ ls
> agg_sta.dat out sa_test2.sh
summary_test.stat
> geeta_test.tar sa_test.sh statfiles
> bash-3.2$
> there were subdirectories which also had the stat files. So I have
removed the subdirectory and Now I am getting only that number of CNT
entries as the No of files. So this problem looks like fixed.
>
> 2. So summary is like collection of the results in one file as
specified by the filter.
> 3. I wish to know that the job aggregate willl take the mean of all
the days output???? or how does it give me a single figure of RMSE for
100 days of verification of RF.
> 4. So If I run the summary job and then extract the RMSE column from
the summary o/p file, and then take a mean, will I arrive at the same
result as that of AGGREGATE????? basically, I wish to know, what
exactly the aggregate does????
> 5. I am sending you the statfiles. Kindly have a look.
> Thanks.
>
> geeta
>
>> Subject: Re: [rt.rap.ucar.edu #57995] Point stat Clarification.
>> From: met_help at ucar.edu
>> To: geeta124 at hotmail.com
>> CC: met_help-owner at mailman.ucar.edu
>> Date: Fri, 24 Aug 2012 09:08:52 -0600
>>
>> Geeta,
>>
>> Your point_stat files may contain more than one CNT line per file.
That depends on how you configured point_stat. For
>> example, verifying multiple forecast variables, thresholds, masking
regions or interpolation types will all cause
>> point_stat to write more than one CNT line to an output file. This
would also explain the appearance of redundant dates
>> in the stat_analysis output.
>>
>> As for your questions about aggregating the MPR lines into a CNT
line, the slow execution is caused by the bootstrapping
>> process. This is an unavoidable drawback to the bootstrapping
process. If you want to increase performance, and you
>> are not interested in the bootstrap confidence intervals, you can
disable this feature by omitting the boot parameters
>> from your stat_analysis command. Although it is difficult to read
in your email, the error that you reported may be
>> caused by the fact that the -boot_seed parameter requires an
argument and you do not appear to have one specified.
>>
>> I hope this answers your questions. If you have any further
questions on this matter, it might be easiest for me to
>> help you if you bundle your point_stat output files into a tarball
and upload it to our FTP site using the following
>> instructions:
>>
>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>
>> Thanks,
>>
>> Paul
>>
>>
>>
>> On 08/24/2012 02:58 AM, Geeta Geeta via RT wrote:
>>>
>>> Fri Aug 24 02:58:57 2012: Request 57995 was acted upon.
>>> Transaction: Ticket created by geeta124 at hotmail.com
>>> Queue: met_help
>>> Subject: Point stat Clarification.
>>> Owner: Nobody
>>> Requestors: geeta124 at hotmail.com
>>> Status: new
>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=57995 >
>>>
>>>
>>>
>>> Hi Paul, I am doing f/c verf of Rainfall for 122 days using MET. I
run WRF for 48hrs FC. 1. I am doing verf of RF using point
observations. I have run point stat tool and now I am trying with
stat-analysis tool. 2. I have to generate the products like ME, MAE,
RMSe etc for my domain for 24 and 48hrs. 3. for this, I have kept 21
stat files in bash-3.2$ cd statfiles/
>>> /oprn/model/wrf3/utils/met/METv3.0/geeta_test/statfiles
>>> bash-3.2$ ls -l
>>> total 3288
>>> -rwxr-xr-x 1 model staff 91922 Jun 11 14:09
point_stat_gt0.0_270000L_20110612_030000V.stat
>>> -rwxr-xr-x 1 model staff 91007 Jun 11 14:09
point_stat_gt0.0_270000L_20110613_030000V.stat
>>> -rwxr-xr-x 1 model staff 91912 Jun 11 14:09
point_stat_gt0.0_270000L_20110615_030000V.stat
>>> -rwxr-xr-x 1 model staff 92219 Jun 11 14:09
point_stat_gt0.0_270000L_20110616_030000V.stat
>>> -rwxr-xr-x 1 model staff 92834 Jun 11 14:09
point_stat_gt0.0_270000L_20110618_030000V.stat
>>> -rwxr-xr-x 1 model staff 90426 Jun 11 14:09
point_stat_gt0.0_270000L_20110621_030000V.stat
>>> -rwxr-xr-x 1 model staff 92539 Jun 11 14:09
point_stat_gt0.0_270000L_20110713_030000V.stat
>>> -rwxr-xr-x 1 model staff 93487 Jun 11 14:09
point_stat_gt0.0_270000L_20110714_030000V.stat
>>> -rwxr-xr-x 1 model staff 214 Jun 11 14:09
point_stat_gt0.0_270000L_20110715_030000V.stat
>>> -rwxr-xr-x 1 model staff 92852 Jun 11 14:09
point_stat_gt0.0_270000L_20110716_030000V.stat
>>> -rwxr-xr-x 1 model staff 92855 Jun 11 14:09
point_stat_gt0.0_270000L_20110719_030000V.stat
>>> -rwxr-xr-x 1 model staff 90045 Jun 11 14:09
point_stat_gt0.0_270000L_20110720_030000V.stat
>>> -rwxr-xr-x 1 model staff 91585 Jun 11 14:09
point_stat_gt0.0_270000L_20110721_030000V.stat
>>> -rwxr-xr-x 1 model staff 89775 Jun 11 14:09
point_stat_gt0.0_270000L_20110723_030000V.stat
>>> -rwxr-xr-x 1 model staff 89765 Jun 11 14:09
point_stat_gt0.0_270000L_20110724_030000V.stat
>>> -rwxr-xr-x 1 model staff 86769 Jun 11 14:09
point_stat_gt0.0_270000L_20110728_030000V.stat
>>> -rwxr-xr-x 1 model staff 88646 Jun 11 14:09
point_stat_gt0.0_270000L_20110729_030000V.stat
>>> -rwxr-xr-x 1 model staff 90156 Jun 11 14:09
point_stat_gt0.0_270000L_20110730_030000V.stat
>>> -rwxr-xr-x 1 model staff 89462 Jun 11 14:09
point_stat_gt0.0_270000L_20110731_030000V.stat
>>> drwxr-xr-x 2 model staff 8192 Jun 14 11:46 tempo
>>> bash-3.2$ ls -l | wc -l
>>> 21
>>>
>>> 4. I have run stat-tool using the
command/oprn/model/wrf3/utils/met/METv3.0/bin/stat_analysis \
>>> -job summary -line_type CNT -column RMSE \
>>> -dump_row summary_test.stat \
>>> -lookin statfiles \
>>> -fcst_var APCP_24 -fcst_lev A24 -v 2 5. It runs fine but I
have following query. In the O/P file, sumary_test_type, I have 214
lines containing CNT. But I have 21 files as input. Also, for example
20110612_03 (one of the dates) is reapeated in the o/p file 6-7 times.
Also the data corresponsing to this day is also different. I thought
that stat-tool extracts all the RMSE columns from the line CNT and
puts them at one place. Pls clarify on this how am I getting so many
lines. I am attaching the O/p file. 6. I had tried with aggregate
tool, but that does not seem to work very efficiently. Processing STAT
file "statfiles/tempo/point_stat_ge7.6_270000L_20110719_030000V.stat"
.... 106 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110720_030000V.stat" ....
107 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110721_030000V.stat" ....
108 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110723_030000V.stat" ....
109 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110724_030000V.stat" ....
110 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110728_030000V.stat" ....
111 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110729_030000V.stat" ....
112 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110730_030000V.stat" ....
113 of 114
>>> Processing STAT file
"statfiles/tempo/point_stat_ge7.6_270000L_20110731_030000V.stat" ....
114 of 114
>>> STAT Lines read = 62698
>>> STAT Lines retained = 62698Processing Job 1: -job aggregate_stat
-line_type MPR -dump_row agg_sta.dat -out_line_type CNT -out_alpha
0.050000 -boot_interval 1 -boot_rep_prop 1.000000 -n_boot_rep 1000
-boot_rng mt19937 -boot_seed -rank_corr_flag 1
>>> Creating STAT output file "agg_sta.dat"
>>> DataLine::get_item(int) -> range check errorbash-3.2$ I have
the aggregate file but it jus shows the MPR lines for each date????
>>> How to proceed further????
>>>
>>> geeta
>>>
>>
>>
>
>
------------------------------------------------
Subject: Point stat Clarification.
From: Paul Oldenburg
Time: Mon Aug 27 09:03:10 2012
JOB_LIST: -job aggregate_stat -line_type MPR -out_line_type CNT
-out_alpha 0.050000 -boot_interval 1 -boot_rep_prop 1.000000
-n_boot_rep 0 -boot_rng mt19937 -boot_seed -rank_corr_flag 1
COL_NAME: TOTAL FBAR FBAR_NCL FBAR_NCU FBAR_BCL FBAR_BCU FSTDEV
FSTDEV_NCL FSTDEV_NCU FSTDEV_BCL FSTDEV_BCU OBAR OBAR_NCL OBAR_NCU
OBAR_BCL OBAR_BCU OSTDEV OSTDEV_NCL OSTDEV_NCU OSTDEV_BCL OSTDEV_BCU
PR_CORR PR_CORR_NCL PR_CORR_NCU PR_CORR_BCL PR_CORR_BCU SP_CORR
KT_CORR RANKS FRANK_TIES ORANK_TIES ME ME_NCL ME_NCU ME_BCL
ME_BCU ESTDEV ESTDEV_NCL ESTDEV_NCU ESTDEV_BCL ESTDEV_BCU MBIAS
MBIAS_BCL MBIAS_BCU MAE MAE_BCL MAE_BCU MSE MSE_BCL MSE_BCU
BCMSE BCMSE_BCL BCMSE_BCU RMSE RMSE_BCL RMSE_BCU E10
E10_BCL E10_BCU E25 E25_BCL E25_BCU E50 E50_BCL E50_BCU E75
E75_BCL E75_BCU E90 E90_BCL E90_BCU
CNT: 11521 13.83580 13.29813 14.37347 NA NA 29.44510
29.06978 29.83031 NA NA 10.56593 10.18274 10.94913
NA NA 20.98544 20.71794 21.25997 NA NA
0.45356 0.43893 0.46794 NA NA 0.41297
0.30047 10714 8183 10148 3.26987 2.77083 3.76890 NA NA
27.32913 26.98078 27.68666 NA NA 1.30947 NA
NA 12.88104 NA NA 757.50848 NA NA 746.81645
NA NA 27.52287 NA NA -13.40020 NA NA
-1.93273 NA NA 0.30015 NA NA 4.40145 NA NA
24.35723 NA NA
------------------------------------------------
More information about the Met_help
mailing list