[Met_help] [rt.rap.ucar.edu #55263] History for Using point_stat tool to verify the rainfall data
Paul Oldenburg via RT
met_help at ucar.edu
Tue Apr 3 08:45:00 MDT 2012
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi Paul,
I have registered a new email account. So I use this new email account to
sent out the question.
Thanks again for your help.
Delia Chen
====================
Hi
I want to verified the rainfall data by using the point_stat.sh.
I have several questions, hope you can help me.
My observation data includes hourly rainfall data, 12hr accumulated
rainfall data and the daily accumulate rainfall data.
My forecast data is the WRF-ARW output data
Below is my processing steps:
Prepare the observation rainfall data in 10 columns ascii format.
For 24hr accumulate rainfall data. The data is excerpted as following.
---------------
ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12 -9999.00
35.5
ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12 -9999.00
80
ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12 -9999.00
14.5
......
---------------
This date set is for 24-hr accumulated rainfall from 20110828 000000UTC to
20110829 000000UTC.
For 12hr accumulate rainfall data. The data is excerpted as following.
---------------
ADPSFC C1I230 20110828_120000 23.9639 120.8372 917 61 12 -9999.00
21.5
ADPSFC 466940 20110828_120000 25.1348 121.7321 27 61 12 -9999.00
29
ADPSFC 466900 20110828_120000 25.1656 121.44 19 61 12 -9999.00 7
......
ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12 -9999.00
14
ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12 -9999.00
51
ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12 -9999.00
7.5
---------------
Using the ascii2nc to produced the observation data in netcdf format.
For 24-hr accumulated rainfall data, the netcdf file is *
precip24_2011082900.nc*
For 12-hr accumulated rainfall data, the netcdf file is *
precip24_20110828_0829_12hr.nc*
Transfer the WRF-ARW output by using UPPV1.0
My output data is saved as one file per hour.
So, it produced several file from UPP program.
for example: wrfprs_d03.000, wrfprs_d03.001, wrfprs_d03.002 .......
wrfprs_d03.024
If I want to verify the first 24hr accumulated rainfall, by the 24-hr
observational accumulated rainfall
The config file was excerpted as following
----------------
fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24" ];
obs_field[] = [ "61/A24" , "61/A24" , "61/A24" , "61/A24" ];
----------------
../bin/point_stat \
../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
../out/ascii2nc/*precip24_2011082800.nc* \
config/PointStat_Rain \
-outdir ../out/point_stat -v 2
If I want to verify the first 24hr accumulated rainfall, by the 12-hr
observational accumulated rainfall
The config file was excerpted as following
----------------
fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24" ];
obs_field[] = [ "61/A12" , "61/A12" , "61/A12" , "61/A12" ];
----------------
../bin/point_stat \
../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
../out/ascii2nc/*precip24_20110828_0829_12hr.nc* \
config/PointStat_Rain \
-outdir ../out/point_stat -v 2
My question is:
Does there have some procedure I am doing wrong?
The step 4 and step 5 should be identical, but the results shows that the
step 5 only compare with 12hr accumulate.
How can I using the 12-hr observational data to verify the 24-hr
accumulated rainfall?
If I want to verify the continues hourly rainfall data with the
observational hourly rainfall data.
How do I put the fcst_file in the execute line? (Since I have several
output file. )
Thank you very much.
--
Delia, Yen-Chu Chen 陳嬿竹
Associate Researcher
Taiwan Typhoon and Flood Research Institute
台灣颱風洪水研究中心 (TTFRI)
Tel : +882-2-2321-9660 ext 143
Email : deliachen at ttfri.narl.org.tw
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #55263] Using point_stat tool to verify the rainfall data
From: Paul Oldenburg
Time: Tue Mar 20 09:02:45 2012
Delia,
Thank you for registering. First of all, I see one potential problem
in the following table in your initial email:
> For 24hr accumulate rainfall data. The data is excerpted as
following.
> ---------------
> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12
-9999.00
> 35.5
For 24 hour accumulation, you should set the "level" column value
(field 8) to 24 instead of 12.
It sounds like you have WRF output that has been post-processed by UPP
into files containing hourly (1hr) accumulated
precip data with the naming format wrfprs_d03.001. If this is the
case, you will need to run the MET tool pcp_combine
to add the accumulations in these files to equal the accumulation
intervals for your observations (12 and 24). This
might be easiest using a command like this, for 24 hour accumulation:
pcp_combine -sum 20110828_00 1 20110829_00 24 -v 3
This command will tell pcp_combine to look in the current directory
for all UPP output GRIB files initialized at
20110828_00 that have 1hr accumulated precip (APCP), sum their precip
values and write it to a NetCDF output file which
contains 24hr APCP. Then, the NetCDF output file can be passed to
point_stat as the model input file.
I hope this helps answer your question. If you still have trouble,
please upload the following files to our FTP server
using the instructions at
http://www.dtcenter.org/met/users/support/met_help.php#ftp and I will
take a look:
* observation ASCII files
* WRF GRIB data post-processed with UPP
* point_stat config file
Thanks,
Paul
On 03/19/2012 09:25 PM, DeliaChen 陳嬿竹 via RT wrote:
>
> Mon Mar 19 21:25:00 2012: Request 55263 was acted upon.
> Transaction: Ticket created by deliachen at ttfri.narl.org.tw
> Queue: met_help
> Subject: Using point_stat tool to verify the rainfall data
> Owner: Nobody
> Requestors: deliachen at ttfri.narl.org.tw
> Status: new
> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263>
>
>
> Hi Paul,
>
> I have registered a new email account. So I use this new email
account to
> sent out the question.
>
> Thanks again for your help.
>
> Delia Chen
>
> ====================
>
> Hi
>
> I want to verified the rainfall data by using the point_stat.sh.
> I have several questions, hope you can help me.
>
> My observation data includes hourly rainfall data, 12hr accumulated
> rainfall data and the daily accumulate rainfall data.
> My forecast data is the WRF-ARW output data
>
>
> Below is my processing steps:
> Prepare the observation rainfall data in 10 columns ascii format.
>
> For 24hr accumulate rainfall data. The data is excerpted as
following.
> ---------------
> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12
-9999.00
> 35.5
> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
> 80
> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
> 14.5
> ......
> ---------------
> This date set is for 24-hr accumulated rainfall from 20110828
000000UTC to
> 20110829 000000UTC.
>
>
> For 12hr accumulate rainfall data. The data is excerpted as
following.
> ---------------
> ADPSFC C1I230 20110828_120000 23.9639 120.8372 917 61 12
-9999.00
> 21.5
> ADPSFC 466940 20110828_120000 25.1348 121.7321 27 61 12
-9999.00
> 29
> ADPSFC 466900 20110828_120000 25.1656 121.44 19 61 12
-9999.00 7
> ......
> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12
-9999.00
> 14
> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
> 51
> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
> 7.5
> ---------------
>
>
> Using the ascii2nc to produced the observation data in netcdf
format.
> For 24-hr accumulated rainfall data, the netcdf file is *
> precip24_2011082900.nc*
> For 12-hr accumulated rainfall data, the netcdf file is *
> precip24_20110828_0829_12hr.nc*
>
> Transfer the WRF-ARW output by using UPPV1.0
> My output data is saved as one file per hour.
> So, it produced several file from UPP program.
>
> for example: wrfprs_d03.000, wrfprs_d03.001, wrfprs_d03.002
.......
> wrfprs_d03.024
>
> If I want to verify the first 24hr accumulated rainfall, by the 24-
hr
> observational accumulated rainfall
>
> The config file was excerpted as following
> ----------------
> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24"
];
> obs_field[] = [ "61/A24" , "61/A24" , "61/A24" , "61/A24" ];
> ----------------
>
> ../bin/point_stat \
> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
> ../out/ascii2nc/*precip24_2011082800.nc* \
> config/PointStat_Rain \
> -outdir ../out/point_stat -v 2
>
> If I want to verify the first 24hr accumulated rainfall, by the 12-
hr
> observational accumulated rainfall
>
> The config file was excerpted as following
> ----------------
> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24"
];
> obs_field[] = [ "61/A12" , "61/A12" , "61/A12" , "61/A12" ];
> ----------------
>
> ../bin/point_stat \
> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
> ../out/ascii2nc/*precip24_20110828_0829_12hr.nc* \
> config/PointStat_Rain \
> -outdir ../out/point_stat -v 2
>
> My question is:
> Does there have some procedure I am doing wrong?
> The step 4 and step 5 should be identical, but the results shows
that the
> step 5 only compare with 12hr accumulate.
> How can I using the 12-hr observational data to verify the 24-hr
> accumulated rainfall?
> If I want to verify the continues hourly rainfall data with the
> observational hourly rainfall data.
> How do I put the fcst_file in the execute line? (Since I have
several
> output file. )
>
> Thank you very much.
>
------------------------------------------------
Subject: Using point_stat tool to verify the rainfall data
From: DeliaChen 陳嬿竹
Time: Wed Mar 21 03:29:07 2012
Dear Paul,
Thank you very much for your reply.
I might have some mistake, when I copy paste the rainfall data.
Right, the "level" column value should be 24.
I have also double check the UPP output file.
The APCP contains the accumulated rainfall for the specific model
hour.
So, I can just used the "wrfprs_d03.024" file to compare with the 24hr
accumulated rainfall data.
According to your email, I have to make sure that the fcst data and
obs
data are within the same accumulated time period before doing the
point_stat. Is that correct?
But if I want to compare the continuous data (for example the RMSE
between
the obs and fcst for the hourly rainfall), how do I prepared the data?
I have upload the sample data to the ftp (chen_data).
Please take a look at it.
Thank you very much.
Best,
Delia Chen
2012/3/20 Paul Oldenburg via RT <met_help at ucar.edu>
> Delia,
>
> Thank you for registering. First of all, I see one potential
problem in
> the following table in your initial email:
>
> > For 24hr accumulate rainfall data. The data is excerpted as
following.
> > ---------------
> > ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61
12
> -9999.00
> > 35.5
>
> For 24 hour accumulation, you should set the "level" column value
(field
> 8) to 24 instead of 12.
>
> It sounds like you have WRF output that has been post-processed by
UPP
> into files containing hourly (1hr) accumulated
> precip data with the naming format wrfprs_d03.001. If this is the
case,
> you will need to run the MET tool pcp_combine
> to add the accumulations in these files to equal the accumulation
> intervals for your observations (12 and 24). This
> might be easiest using a command like this, for 24 hour
accumulation:
>
> pcp_combine -sum 20110828_00 1 20110829_00 24 -v 3
>
> This command will tell pcp_combine to look in the current directory
for
> all UPP output GRIB files initialized at
> 20110828_00 that have 1hr accumulated precip (APCP), sum their
precip
> values and write it to a NetCDF output file which
> contains 24hr APCP. Then, the NetCDF output file can be passed to
> point_stat as the model input file.
>
> I hope this helps answer your question. If you still have trouble,
please
> upload the following files to our FTP server
> using the instructions at
> http://www.dtcenter.org/met/users/support/met_help.php#ftp and I
will
> take a look:
>
> * observation ASCII files
> * WRF GRIB data post-processed with UPP
> * point_stat config file
>
> Thanks,
>
> Paul
>
>
> On 03/19/2012 09:25 PM, DeliaChen 陳嬿竹 via RT wrote:
> >
> > Mon Mar 19 21:25:00 2012: Request 55263 was acted upon.
> > Transaction: Ticket created by deliachen at ttfri.narl.org.tw
> > Queue: met_help
> > Subject: Using point_stat tool to verify the rainfall data
> > Owner: Nobody
> > Requestors: deliachen at ttfri.narl.org.tw
> > Status: new
> > Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263>
> >
> >
> > Hi Paul,
> >
> > I have registered a new email account. So I use this new email
account to
> > sent out the question.
> >
> > Thanks again for your help.
> >
> > Delia Chen
> >
> > ====================
> >
> > Hi
> >
> > I want to verified the rainfall data by using the point_stat.sh.
> > I have several questions, hope you can help me.
> >
> > My observation data includes hourly rainfall data, 12hr
accumulated
> > rainfall data and the daily accumulate rainfall data.
> > My forecast data is the WRF-ARW output data
> >
> >
> > Below is my processing steps:
> > Prepare the observation rainfall data in 10 columns ascii format.
> >
> > For 24hr accumulate rainfall data. The data is excerpted as
following.
> > ---------------
> > ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12
> -9999.00
> > 35.5
> > ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
> > 80
> > ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
> > 14.5
> > ......
> > ---------------
> > This date set is for 24-hr accumulated rainfall from 20110828
000000UTC
> to
> > 20110829 000000UTC.
> >
> >
> > For 12hr accumulate rainfall data. The data is excerpted as
following.
> > ---------------
> > ADPSFC C1I230 20110828_120000 23.9639 120.8372 917 61 12
> -9999.00
> > 21.5
> > ADPSFC 466940 20110828_120000 25.1348 121.7321 27 61 12
-9999.00
> > 29
> > ADPSFC 466900 20110828_120000 25.1656 121.44 19 61 12
-9999.00
> 7
> > ......
> > ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12
> -9999.00
> > 14
> > ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
> > 51
> > ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
> > 7.5
> > ---------------
> >
> >
> > Using the ascii2nc to produced the observation data in netcdf
format.
> > For 24-hr accumulated rainfall data, the netcdf file is *
> > precip24_2011082900.nc*
> > For 12-hr accumulated rainfall data, the netcdf file is *
> > precip24_20110828_0829_12hr.nc*
> >
> > Transfer the WRF-ARW output by using UPPV1.0
> > My output data is saved as one file per hour.
> > So, it produced several file from UPP program.
> >
> > for example: wrfprs_d03.000, wrfprs_d03.001, wrfprs_d03.002
.......
> > wrfprs_d03.024
> >
> > If I want to verify the first 24hr accumulated rainfall, by the
24-hr
> > observational accumulated rainfall
> >
> > The config file was excerpted as following
> > ----------------
> > fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24"
];
> > obs_field[] = [ "61/A24" , "61/A24" , "61/A24" , "61/A24" ];
> > ----------------
> >
> > ../bin/point_stat \
> > ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
> > ../out/ascii2nc/*precip24_2011082800.nc* \
> > config/PointStat_Rain \
> > -outdir ../out/point_stat -v 2
> >
> > If I want to verify the first 24hr accumulated rainfall, by the
12-hr
> > observational accumulated rainfall
> >
> > The config file was excerpted as following
> > ----------------
> > fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24"
];
> > obs_field[] = [ "61/A12" , "61/A12" , "61/A12" , "61/A12" ];
> > ----------------
> >
> > ../bin/point_stat \
> > ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
> > ../out/ascii2nc/*precip24_20110828_0829_12hr.nc* \
> > config/PointStat_Rain \
> > -outdir ../out/point_stat -v 2
> >
> > My question is:
> > Does there have some procedure I am doing wrong?
> > The step 4 and step 5 should be identical, but the results shows
that the
> > step 5 only compare with 12hr accumulate.
> > How can I using the 12-hr observational data to verify the 24-hr
> > accumulated rainfall?
> > If I want to verify the continues hourly rainfall data with the
> > observational hourly rainfall data.
> > How do I put the fcst_file in the execute line? (Since I have
several
> > output file. )
> >
> > Thank you very much.
> >
>
>
>
--
Delia, Yen-Chu Chen 陳嬿竹
Associate Researcher
Taiwan Typhoon and Flood Research Institute
台灣颱風洪水研究中心 (TTFRI)
Tel : +882-2-2321-9660 ext 143
Email : deliachen at ttfri.narl.org.tw
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #55263] Using point_stat tool to verify the rainfall data
From: Paul Oldenburg
Time: Wed Mar 21 08:55:52 2012
Delia,
> According to your email, I have to make sure that the fcst data and
obs
> data are within the same accumulated time period before doing the
> point_stat. Is that correct?
This is correct, which is why you will need hourly 1hr accumulated
rainfall observations to verify the hourly model
accumulated rainfall predictions. I checked your observation text
file, 20110828_29.dat, and it appears to only contain
24 hour precip obs. Do you have similar obs for hourly 1hr precip?
If you are interested in generating continuous statistics from any
verification comparison in point_stat (or other MET
tools), set the output_flag setting in the configuration file so that
the 6th and 7th flags are either 1 or 2. Then,
point_stat will write out many continuous statistics calculated during
verification.
Paul
On 03/21/2012 03:29 AM, DeliaChen 陳嬿竹 via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263>
>
> Dear Paul,
>
> Thank you very much for your reply.
>
> I might have some mistake, when I copy paste the rainfall data.
> Right, the "level" column value should be 24.
>
> I have also double check the UPP output file.
> The APCP contains the accumulated rainfall for the specific model
hour.
> So, I can just used the "wrfprs_d03.024" file to compare with the
24hr
> accumulated rainfall data.
>
> According to your email, I have to make sure that the fcst data and
obs
> data are within the same accumulated time period before doing the
> point_stat. Is that correct?
>
> But if I want to compare the continuous data (for example the RMSE
between
> the obs and fcst for the hourly rainfall), how do I prepared the
data?
> I have upload the sample data to the ftp (chen_data).
> Please take a look at it.
>
> Thank you very much.
> Best,
>
> Delia Chen
>
>
> 2012/3/20 Paul Oldenburg via RT<met_help at ucar.edu>
>
>> Delia,
>>
>> Thank you for registering. First of all, I see one potential
problem in
>> the following table in your initial email:
>>
>> > For 24hr accumulate rainfall data. The data is excerpted as
following.
>> > ---------------
>> > ADPSFC C1I230 20110829_000000 23.9639 120.8372 917
61 12
>> -9999.00
>> > 35.5
>>
>> For 24 hour accumulation, you should set the "level" column value
(field
>> 8) to 24 instead of 12.
>>
>> It sounds like you have WRF output that has been post-processed by
UPP
>> into files containing hourly (1hr) accumulated
>> precip data with the naming format wrfprs_d03.001. If this is the
case,
>> you will need to run the MET tool pcp_combine
>> to add the accumulations in these files to equal the accumulation
>> intervals for your observations (12 and 24). This
>> might be easiest using a command like this, for 24 hour
accumulation:
>>
>> pcp_combine -sum 20110828_00 1 20110829_00 24 -v 3
>>
>> This command will tell pcp_combine to look in the current directory
for
>> all UPP output GRIB files initialized at
>> 20110828_00 that have 1hr accumulated precip (APCP), sum their
precip
>> values and write it to a NetCDF output file which
>> contains 24hr APCP. Then, the NetCDF output file can be passed to
>> point_stat as the model input file.
>>
>> I hope this helps answer your question. If you still have trouble,
please
>> upload the following files to our FTP server
>> using the instructions at
>> http://www.dtcenter.org/met/users/support/met_help.php#ftp and I
will
>> take a look:
>>
>> * observation ASCII files
>> * WRF GRIB data post-processed with UPP
>> * point_stat config file
>>
>> Thanks,
>>
>> Paul
>>
>>
>> On 03/19/2012 09:25 PM, DeliaChen 陳嬿竹 via RT wrote:
>>>
>>> Mon Mar 19 21:25:00 2012: Request 55263 was acted upon.
>>> Transaction: Ticket created by deliachen at ttfri.narl.org.tw
>>> Queue: met_help
>>> Subject: Using point_stat tool to verify the rainfall data
>>> Owner: Nobody
>>> Requestors: deliachen at ttfri.narl.org.tw
>>> Status: new
>>> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263>
>>>
>>>
>>> Hi Paul,
>>>
>>> I have registered a new email account. So I use this new email
account to
>>> sent out the question.
>>>
>>> Thanks again for your help.
>>>
>>> Delia Chen
>>>
>>> ====================
>>>
>>> Hi
>>>
>>> I want to verified the rainfall data by using the point_stat.sh.
>>> I have several questions, hope you can help me.
>>>
>>> My observation data includes hourly rainfall data, 12hr
accumulated
>>> rainfall data and the daily accumulate rainfall data.
>>> My forecast data is the WRF-ARW output data
>>>
>>>
>>> Below is my processing steps:
>>> Prepare the observation rainfall data in 10 columns ascii format.
>>>
>>> For 24hr accumulate rainfall data. The data is excerpted as
following.
>>> ---------------
>>> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12
>> -9999.00
>>> 35.5
>>> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
>>> 80
>>> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
>>> 14.5
>>> ......
>>> ---------------
>>> This date set is for 24-hr accumulated rainfall from 20110828
000000UTC
>> to
>>> 20110829 000000UTC.
>>>
>>>
>>> For 12hr accumulate rainfall data. The data is excerpted as
following.
>>> ---------------
>>> ADPSFC C1I230 20110828_120000 23.9639 120.8372 917 61 12
>> -9999.00
>>> 21.5
>>> ADPSFC 466940 20110828_120000 25.1348 121.7321 27 61 12
-9999.00
>>> 29
>>> ADPSFC 466900 20110828_120000 25.1656 121.44 19 61 12
-9999.00
>> 7
>>> ......
>>> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61 12
>> -9999.00
>>> 14
>>> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
>>> 51
>>> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
>>> 7.5
>>> ---------------
>>>
>>>
>>> Using the ascii2nc to produced the observation data in netcdf
format.
>>> For 24-hr accumulated rainfall data, the netcdf file is *
>>> precip24_2011082900.nc*
>>> For 12-hr accumulated rainfall data, the netcdf file is *
>>> precip24_20110828_0829_12hr.nc*
>>>
>>> Transfer the WRF-ARW output by using UPPV1.0
>>> My output data is saved as one file per hour.
>>> So, it produced several file from UPP program.
>>>
>>> for example: wrfprs_d03.000, wrfprs_d03.001, wrfprs_d03.002
.......
>>> wrfprs_d03.024
>>>
>>> If I want to verify the first 24hr accumulated rainfall, by the
24-hr
>>> observational accumulated rainfall
>>>
>>> The config file was excerpted as following
>>> ----------------
>>> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24"
];
>>> obs_field[] = [ "61/A24" , "61/A24" , "61/A24" , "61/A24" ];
>>> ----------------
>>>
>>> ../bin/point_stat \
>>> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
>>> ../out/ascii2nc/*precip24_2011082800.nc* \
>>> config/PointStat_Rain \
>>> -outdir ../out/point_stat -v 2
>>>
>>> If I want to verify the first 24hr accumulated rainfall, by the
12-hr
>>> observational accumulated rainfall
>>>
>>> The config file was excerpted as following
>>> ----------------
>>> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" , "APCP/A24"
];
>>> obs_field[] = [ "61/A12" , "61/A12" , "61/A12" , "61/A12" ];
>>> ----------------
>>>
>>> ../bin/point_stat \
>>> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
>>> ../out/ascii2nc/*precip24_20110828_0829_12hr.nc* \
>>> config/PointStat_Rain \
>>> -outdir ../out/point_stat -v 2
>>>
>>> My question is:
>>> Does there have some procedure I am doing wrong?
>>> The step 4 and step 5 should be identical, but the results shows
that the
>>> step 5 only compare with 12hr accumulate.
>>> How can I using the 12-hr observational data to verify the 24-hr
>>> accumulated rainfall?
>>> If I want to verify the continues hourly rainfall data with the
>>> observational hourly rainfall data.
>>> How do I put the fcst_file in the execute line? (Since I have
several
>>> output file. )
>>>
>>> Thank you very much.
>>>
>>
>>
>>
>
>
------------------------------------------------
Subject: Using point_stat tool to verify the rainfall data
From: DeliaChen 陳嬿竹
Time: Wed Mar 21 18:33:31 2012
Hi Paul,
Yes, I have the obs for hourly 1hr precip, I might be upload the wrong
data.
I hvae also set the output_flag for it.
My problem is how do I put the fcst data in the command line?
../bin/point_stat \
../../POST_UPP/M15_2011082800/*wrfprs_d03.006* \
../out/ascii2nc/precip_20110828_0829_1hr.nc \
config/PointStat_Rain \
-outdir ../out/point_stat -v 2
I have already check that the precipitation output from UPP is the
accumulated rainfall, and the output is 1hr per file.
Should I need to transform the accumulated rainfall to hourly
rainfall, and
then combined all the data to become one file? Thus I used this file
(contains several hourly rainfall) to do the verification.
Is this correct?
Thank you very much. It's become more clearly now.
Best,
Delia Chen
2012/3/21 Paul Oldenburg via RT <met_help at ucar.edu>
> Delia,
>
> > According to your email, I have to make sure that the fcst data
and obs
> > data are within the same accumulated time period before doing the
> > point_stat. Is that correct?
>
> This is correct, which is why you will need hourly 1hr accumulated
> rainfall observations to verify the hourly model
> accumulated rainfall predictions. I checked your observation text
file,
> 20110828_29.dat, and it appears to only contain
> 24 hour precip obs. Do you have similar obs for hourly 1hr precip?
>
> If you are interested in generating continuous statistics from any
> verification comparison in point_stat (or other MET
> tools), set the output_flag setting in the configuration file so
that the
> 6th and 7th flags are either 1 or 2. Then,
> point_stat will write out many continuous statistics calculated
during
> verification.
>
> Paul
>
>
> On 03/21/2012 03:29 AM, DeliaChen 陳嬿竹 via RT wrote:
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263>
> >
> > Dear Paul,
> >
> > Thank you very much for your reply.
> >
> > I might have some mistake, when I copy paste the rainfall data.
> > Right, the "level" column value should be 24.
> >
> > I have also double check the UPP output file.
> > The APCP contains the accumulated rainfall for the specific model
hour.
> > So, I can just used the "wrfprs_d03.024" file to compare with the
24hr
> > accumulated rainfall data.
> >
> > According to your email, I have to make sure that the fcst data
and obs
> > data are within the same accumulated time period before doing the
> > point_stat. Is that correct?
> >
> > But if I want to compare the continuous data (for example the RMSE
> between
> > the obs and fcst for the hourly rainfall), how do I prepared the
data?
> > I have upload the sample data to the ftp (chen_data).
> > Please take a look at it.
> >
> > Thank you very much.
> > Best,
> >
> > Delia Chen
> >
> >
> > 2012/3/20 Paul Oldenburg via RT<met_help at ucar.edu>
> >
> >> Delia,
> >>
> >> Thank you for registering. First of all, I see one potential
problem in
> >> the following table in your initial email:
> >>
> >> > For 24hr accumulate rainfall data. The data is excerpted as
> following.
> >> > ---------------
> >> > ADPSFC C1I230 20110829_000000 23.9639 120.8372 917
61 12
> >> -9999.00
> >> > 35.5
> >>
> >> For 24 hour accumulation, you should set the "level" column value
(field
> >> 8) to 24 instead of 12.
> >>
> >> It sounds like you have WRF output that has been post-processed
by UPP
> >> into files containing hourly (1hr) accumulated
> >> precip data with the naming format wrfprs_d03.001. If this is
the case,
> >> you will need to run the MET tool pcp_combine
> >> to add the accumulations in these files to equal the accumulation
> >> intervals for your observations (12 and 24). This
> >> might be easiest using a command like this, for 24 hour
accumulation:
> >>
> >> pcp_combine -sum 20110828_00 1 20110829_00 24 -v 3
> >>
> >> This command will tell pcp_combine to look in the current
directory for
> >> all UPP output GRIB files initialized at
> >> 20110828_00 that have 1hr accumulated precip (APCP), sum their
precip
> >> values and write it to a NetCDF output file which
> >> contains 24hr APCP. Then, the NetCDF output file can be passed
to
> >> point_stat as the model input file.
> >>
> >> I hope this helps answer your question. If you still have
trouble,
> please
> >> upload the following files to our FTP server
> >> using the instructions at
> >> http://www.dtcenter.org/met/users/support/met_help.php#ftp and I
will
> >> take a look:
> >>
> >> * observation ASCII files
> >> * WRF GRIB data post-processed with UPP
> >> * point_stat config file
> >>
> >> Thanks,
> >>
> >> Paul
> >>
> >>
> >> On 03/19/2012 09:25 PM, DeliaChen 陳嬿竹 via RT wrote:
> >>>
> >>> Mon Mar 19 21:25:00 2012: Request 55263 was acted upon.
> >>> Transaction: Ticket created by deliachen at ttfri.narl.org.tw
> >>> Queue: met_help
> >>> Subject: Using point_stat tool to verify the rainfall
data
> >>> Owner: Nobody
> >>> Requestors: deliachen at ttfri.narl.org.tw
> >>> Status: new
> >>> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263
> >
> >>>
> >>>
> >>> Hi Paul,
> >>>
> >>> I have registered a new email account. So I use this new email
account
> to
> >>> sent out the question.
> >>>
> >>> Thanks again for your help.
> >>>
> >>> Delia Chen
> >>>
> >>> ====================
> >>>
> >>> Hi
> >>>
> >>> I want to verified the rainfall data by using the point_stat.sh.
> >>> I have several questions, hope you can help me.
> >>>
> >>> My observation data includes hourly rainfall data, 12hr
accumulated
> >>> rainfall data and the daily accumulate rainfall data.
> >>> My forecast data is the WRF-ARW output data
> >>>
> >>>
> >>> Below is my processing steps:
> >>> Prepare the observation rainfall data in 10 columns ascii
format.
> >>>
> >>> For 24hr accumulate rainfall data. The data is excerpted as
following.
> >>> ---------------
> >>> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61
12
> >> -9999.00
> >>> 35.5
> >>> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
> -9999.00
> >>> 80
> >>> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
> -9999.00
> >>> 14.5
> >>> ......
> >>> ---------------
> >>> This date set is for 24-hr accumulated rainfall from 20110828
000000UTC
> >> to
> >>> 20110829 000000UTC.
> >>>
> >>>
> >>> For 12hr accumulate rainfall data. The data is excerpted as
following.
> >>> ---------------
> >>> ADPSFC C1I230 20110828_120000 23.9639 120.8372 917 61
12
> >> -9999.00
> >>> 21.5
> >>> ADPSFC 466940 20110828_120000 25.1348 121.7321 27 61 12
-9999.00
> >>> 29
> >>> ADPSFC 466900 20110828_120000 25.1656 121.44 19 61 12
-9999.00
> >> 7
> >>> ......
> >>> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61
12
> >> -9999.00
> >>> 14
> >>> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
> >>> 51
> >>> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
> >>> 7.5
> >>> ---------------
> >>>
> >>>
> >>> Using the ascii2nc to produced the observation data in netcdf
format.
> >>> For 24-hr accumulated rainfall data, the netcdf file is *
> >>> precip24_2011082900.nc*
> >>> For 12-hr accumulated rainfall data, the netcdf file is *
> >>> precip24_20110828_0829_12hr.nc*
> >>>
> >>> Transfer the WRF-ARW output by using UPPV1.0
> >>> My output data is saved as one file per hour.
> >>> So, it produced several file from UPP program.
> >>>
> >>> for example: wrfprs_d03.000, wrfprs_d03.001, wrfprs_d03.002
.......
> >>> wrfprs_d03.024
> >>>
> >>> If I want to verify the first 24hr accumulated rainfall, by the
24-hr
> >>> observational accumulated rainfall
> >>>
> >>> The config file was excerpted as following
> >>> ----------------
> >>> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" ,
"APCP/A24" ];
> >>> obs_field[] = [ "61/A24" , "61/A24" , "61/A24" , "61/A24" ];
> >>> ----------------
> >>>
> >>> ../bin/point_stat \
> >>> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
> >>> ../out/ascii2nc/*precip24_2011082800.nc* \
> >>> config/PointStat_Rain \
> >>> -outdir ../out/point_stat -v 2
> >>>
> >>> If I want to verify the first 24hr accumulated rainfall, by the
12-hr
> >>> observational accumulated rainfall
> >>>
> >>> The config file was excerpted as following
> >>> ----------------
> >>> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" ,
"APCP/A24" ];
> >>> obs_field[] = [ "61/A12" , "61/A12" , "61/A12" , "61/A12" ];
> >>> ----------------
> >>>
> >>> ../bin/point_stat \
> >>> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
> >>> ../out/ascii2nc/*precip24_20110828_0829_12hr.nc* \
> >>> config/PointStat_Rain \
> >>> -outdir ../out/point_stat -v 2
> >>>
> >>> My question is:
> >>> Does there have some procedure I am doing wrong?
> >>> The step 4 and step 5 should be identical, but the results shows
that
> the
> >>> step 5 only compare with 12hr accumulate.
> >>> How can I using the 12-hr observational data to verify the 24-hr
> >>> accumulated rainfall?
> >>> If I want to verify the continues hourly rainfall data with the
> >>> observational hourly rainfall data.
> >>> How do I put the fcst_file in the execute line? (Since I have
several
> >>> output file. )
> >>>
> >>> Thank you very much.
> >>>
> >>
> >>
> >>
> >
> >
>
>
>
--
Delia, Yen-Chu Chen 陳嬿竹
Associate Researcher
Taiwan Typhoon and Flood Research Institute
台灣颱風洪水研究中心 (TTFRI)
Tel : +882-2-2321-9660 ext 143
Email : deliachen at ttfri.narl.org.tw
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #55263] Using point_stat tool to verify the rainfall data
From: Paul Oldenburg
Time: Wed Mar 21 21:46:18 2012
Delia,
First of all, it looks like your model data has what we call "run-
time" accumulation. In other words, the file for lead
time N contains the precip accumulation interval 0-N. For example,
the file WRFPRS_d03.003 contains accumulation data
for 0-3:
[pgoldenb at orval 20120321.chen]$ wgrib WRFPRS_d03.003 | grep APCP
278:11484694:d=11082800:APCP:kpds5=61:kpds6=1:kpds7=0:TR=4:P1=0:P2=3:TimeU=1:sfc:0-
3hr acc:NAve=0
To calculate hourly 1hr precip for "run-time" accumulation files, use
the MET tool pcp_combine to subtract the
accumulation for lead time N-1 from the accumulation for lead time N:
$ $MET_base/BIN/pcp_combine -subtract WRFPRS_d03.003 3 WRFPRS_d03.002
2 WRFPRS_d03.003.nc
You may want to write a script that will perform this calculation on
all of your model data files with lead times 2 or
greater.
You will also need hourly 1hr accumulated precip observations. I see
that the accumulations in the file
precip_20110828_0829_1hr.nc are actually 24 hour (86400 sec). Was
this file generated from the text file
20110828_29.dat? I used the commands below to see this:
[pgoldenb at orval 20120321.chen]$ ncdump -h precip_20110828_0829_1hr.nc
...
float obs_arr(nobs, obs_arr_len) ;
obs_arr:long_name = "array of observation values" ;
obs_arr:_fill_value = -9999.f ;
obs_arr:columns = "hdr_id gc lvl hgt ob" ;
$ ncdump -v obs_arr precip_20110828_0829_1hr.nc | egrep '^ +[0-9]' |
head
0, 61, 86400, -9999, -9999,
1, 61, 86400, -9999, -9999,
...
Once you have both 1hr hourly accumulations for both model and obs,
you will need to call point_stat once per model data
file. Unfortunately, MET tools including point_stat do not work with
multiple input model files. You may want to
script this process as well. Sorry if this sounds complicated. If
you have trouble, upload some 1hr obs text data to
the FTP site, and I will take a look.
Good luck,
Paul
On 03/21/2012 06:33 PM, DeliaChen 陳嬿竹 via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263>
>
> Hi Paul,
>
> Yes, I have the obs for hourly 1hr precip, I might be upload the
wrong data.
> I hvae also set the output_flag for it.
>
> My problem is how do I put the fcst data in the command line?
>
> ../bin/point_stat \
> ../../POST_UPP/M15_2011082800/*wrfprs_d03.006* \
> ../out/ascii2nc/precip_20110828_0829_1hr.nc \
> config/PointStat_Rain \
> -outdir ../out/point_stat -v 2
>
> I have already check that the precipitation output from UPP is the
> accumulated rainfall, and the output is 1hr per file.
> Should I need to transform the accumulated rainfall to hourly
rainfall, and
> then combined all the data to become one file? Thus I used this file
> (contains several hourly rainfall) to do the verification.
> Is this correct?
>
> Thank you very much. It's become more clearly now.
>
> Best,
> Delia Chen
>
>
>
> 2012/3/21 Paul Oldenburg via RT<met_help at ucar.edu>
>
>> Delia,
>>
>> > According to your email, I have to make sure that the fcst
data and obs
>> > data are within the same accumulated time period before doing
the
>> > point_stat. Is that correct?
>>
>> This is correct, which is why you will need hourly 1hr accumulated
>> rainfall observations to verify the hourly model
>> accumulated rainfall predictions. I checked your observation text
file,
>> 20110828_29.dat, and it appears to only contain
>> 24 hour precip obs. Do you have similar obs for hourly 1hr precip?
>>
>> If you are interested in generating continuous statistics from any
>> verification comparison in point_stat (or other MET
>> tools), set the output_flag setting in the configuration file so
that the
>> 6th and 7th flags are either 1 or 2. Then,
>> point_stat will write out many continuous statistics calculated
during
>> verification.
>>
>> Paul
>>
>>
>> On 03/21/2012 03:29 AM, DeliaChen 陳嬿竹 via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263>
>>>
>>> Dear Paul,
>>>
>>> Thank you very much for your reply.
>>>
>>> I might have some mistake, when I copy paste the rainfall data.
>>> Right, the "level" column value should be 24.
>>>
>>> I have also double check the UPP output file.
>>> The APCP contains the accumulated rainfall for the specific model
hour.
>>> So, I can just used the "wrfprs_d03.024" file to compare with the
24hr
>>> accumulated rainfall data.
>>>
>>> According to your email, I have to make sure that the fcst data
and obs
>>> data are within the same accumulated time period before doing the
>>> point_stat. Is that correct?
>>>
>>> But if I want to compare the continuous data (for example the RMSE
>> between
>>> the obs and fcst for the hourly rainfall), how do I prepared the
data?
>>> I have upload the sample data to the ftp (chen_data).
>>> Please take a look at it.
>>>
>>> Thank you very much.
>>> Best,
>>>
>>> Delia Chen
>>>
>>>
>>> 2012/3/20 Paul Oldenburg via RT<met_help at ucar.edu>
>>>
>>>> Delia,
>>>>
>>>> Thank you for registering. First of all, I see one potential
problem in
>>>> the following table in your initial email:
>>>>
>>>> > For 24hr accumulate rainfall data. The data is excerpted
as
>> following.
>>>> > ---------------
>>>> > ADPSFC C1I230 20110829_000000 23.9639 120.8372 917
61 12
>>>> -9999.00
>>>> > 35.5
>>>>
>>>> For 24 hour accumulation, you should set the "level" column value
(field
>>>> 8) to 24 instead of 12.
>>>>
>>>> It sounds like you have WRF output that has been post-processed
by UPP
>>>> into files containing hourly (1hr) accumulated
>>>> precip data with the naming format wrfprs_d03.001. If this is
the case,
>>>> you will need to run the MET tool pcp_combine
>>>> to add the accumulations in these files to equal the accumulation
>>>> intervals for your observations (12 and 24). This
>>>> might be easiest using a command like this, for 24 hour
accumulation:
>>>>
>>>> pcp_combine -sum 20110828_00 1 20110829_00 24 -v 3
>>>>
>>>> This command will tell pcp_combine to look in the current
directory for
>>>> all UPP output GRIB files initialized at
>>>> 20110828_00 that have 1hr accumulated precip (APCP), sum their
precip
>>>> values and write it to a NetCDF output file which
>>>> contains 24hr APCP. Then, the NetCDF output file can be passed
to
>>>> point_stat as the model input file.
>>>>
>>>> I hope this helps answer your question. If you still have
trouble,
>> please
>>>> upload the following files to our FTP server
>>>> using the instructions at
>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp and I
will
>>>> take a look:
>>>>
>>>> * observation ASCII files
>>>> * WRF GRIB data post-processed with UPP
>>>> * point_stat config file
>>>>
>>>> Thanks,
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 03/19/2012 09:25 PM, DeliaChen 陳嬿竹 via RT wrote:
>>>>>
>>>>> Mon Mar 19 21:25:00 2012: Request 55263 was acted upon.
>>>>> Transaction: Ticket created by deliachen at ttfri.narl.org.tw
>>>>> Queue: met_help
>>>>> Subject: Using point_stat tool to verify the rainfall
data
>>>>> Owner: Nobody
>>>>> Requestors: deliachen at ttfri.narl.org.tw
>>>>> Status: new
>>>>> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=55263
>>>
>>>>>
>>>>>
>>>>> Hi Paul,
>>>>>
>>>>> I have registered a new email account. So I use this new email
account
>> to
>>>>> sent out the question.
>>>>>
>>>>> Thanks again for your help.
>>>>>
>>>>> Delia Chen
>>>>>
>>>>> ====================
>>>>>
>>>>> Hi
>>>>>
>>>>> I want to verified the rainfall data by using the point_stat.sh.
>>>>> I have several questions, hope you can help me.
>>>>>
>>>>> My observation data includes hourly rainfall data, 12hr
accumulated
>>>>> rainfall data and the daily accumulate rainfall data.
>>>>> My forecast data is the WRF-ARW output data
>>>>>
>>>>>
>>>>> Below is my processing steps:
>>>>> Prepare the observation rainfall data in 10 columns ascii
format.
>>>>>
>>>>> For 24hr accumulate rainfall data. The data is excerpted as
following.
>>>>> ---------------
>>>>> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61
12
>>>> -9999.00
>>>>> 35.5
>>>>> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
>> -9999.00
>>>>> 80
>>>>> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
>> -9999.00
>>>>> 14.5
>>>>> ......
>>>>> ---------------
>>>>> This date set is for 24-hr accumulated rainfall from 20110828
000000UTC
>>>> to
>>>>> 20110829 000000UTC.
>>>>>
>>>>>
>>>>> For 12hr accumulate rainfall data. The data is excerpted as
following.
>>>>> ---------------
>>>>> ADPSFC C1I230 20110828_120000 23.9639 120.8372 917 61
12
>>>> -9999.00
>>>>> 21.5
>>>>> ADPSFC 466940 20110828_120000 25.1348 121.7321 27 61 12
-9999.00
>>>>> 29
>>>>> ADPSFC 466900 20110828_120000 25.1656 121.44 19 61 12
-9999.00
>>>> 7
>>>>> ......
>>>>> ADPSFC C1I230 20110829_000000 23.9639 120.8372 917 61
12
>>>> -9999.00
>>>>> 14
>>>>> ADPSFC 466940 20110829_000000 25.1348 121.7321 27 61 12
-9999.00
>>>>> 51
>>>>> ADPSFC 466900 20110829_000000 25.1656 121.44 19 61 12
-9999.00
>>>>> 7.5
>>>>> ---------------
>>>>>
>>>>>
>>>>> Using the ascii2nc to produced the observation data in netcdf
format.
>>>>> For 24-hr accumulated rainfall data, the netcdf file is *
>>>>> precip24_2011082900.nc*
>>>>> For 12-hr accumulated rainfall data, the netcdf file is *
>>>>> precip24_20110828_0829_12hr.nc*
>>>>>
>>>>> Transfer the WRF-ARW output by using UPPV1.0
>>>>> My output data is saved as one file per hour.
>>>>> So, it produced several file from UPP program.
>>>>>
>>>>> for example: wrfprs_d03.000, wrfprs_d03.001, wrfprs_d03.002
.......
>>>>> wrfprs_d03.024
>>>>>
>>>>> If I want to verify the first 24hr accumulated rainfall, by the
24-hr
>>>>> observational accumulated rainfall
>>>>>
>>>>> The config file was excerpted as following
>>>>> ----------------
>>>>> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" ,
"APCP/A24" ];
>>>>> obs_field[] = [ "61/A24" , "61/A24" , "61/A24" , "61/A24" ];
>>>>> ----------------
>>>>>
>>>>> ../bin/point_stat \
>>>>> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
>>>>> ../out/ascii2nc/*precip24_2011082800.nc* \
>>>>> config/PointStat_Rain \
>>>>> -outdir ../out/point_stat -v 2
>>>>>
>>>>> If I want to verify the first 24hr accumulated rainfall, by the
12-hr
>>>>> observational accumulated rainfall
>>>>>
>>>>> The config file was excerpted as following
>>>>> ----------------
>>>>> fcst_field[] = [ "APCP/A24" , "APCP/A24" , "APCP/A24" ,
"APCP/A24" ];
>>>>> obs_field[] = [ "61/A12" , "61/A12" , "61/A12" , "61/A12" ];
>>>>> ----------------
>>>>>
>>>>> ../bin/point_stat \
>>>>> ../../POST_UPP/M15_2011082800/wrfprs_d03.024 \
>>>>> ../out/ascii2nc/*precip24_20110828_0829_12hr.nc* \
>>>>> config/PointStat_Rain \
>>>>> -outdir ../out/point_stat -v 2
>>>>>
>>>>> My question is:
>>>>> Does there have some procedure I am doing wrong?
>>>>> The step 4 and step 5 should be identical, but the results shows
that
>> the
>>>>> step 5 only compare with 12hr accumulate.
>>>>> How can I using the 12-hr observational data to verify the 24-hr
>>>>> accumulated rainfall?
>>>>> If I want to verify the continues hourly rainfall data with the
>>>>> observational hourly rainfall data.
>>>>> How do I put the fcst_file in the execute line? (Since I have
several
>>>>> output file. )
>>>>>
>>>>> Thank you very much.
>>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>>
>
>
------------------------------------------------
More information about the Met_help
mailing list