[Met_help] [rt.rap.ucar.edu #47600] History for point stat question problem
RAL HelpDesk {for Paul Oldenburg}
met_help at ucar.edu
Fri Jun 17 13:00:10 MDT 2011
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Paul et al,
I noticed that for QVAPOR and Q2 debiased square root error is zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug or something wrong with my processing.
Thanks,
Mariusz
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: point stat question problem
From: Paul Oldenburg
Time: Thu Jun 16 16:48:12 2011
Mariusz,
You have uncovered another issue in MET related to precision. Your
data set certainly is testing our numerical
thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
file and place the attached version in it's place. Then, run make
clean in the lib/vx_math folder, rebuild MET from the
top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
square root error. Sorry for the trouble. Please let me know if you
have any questions.
Paul
On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>
> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
> Queue: met_help
> Subject: point stat question problem
> Owner: Nobody
> Requestors: Mariusz.Pagowski at noaa.gov
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>
>
> Paul et al,
> I noticed that for QVAPOR and Q2 debiased square root error is zero
in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug or
something wrong with my processing.
> Thanks,
> Mariusz
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Mariusz Pagowski
Time: Thu Jun 16 21:46:33 2011
Paul,
the same problem with Mac e-mail. Can you send the file separately,
not inside the text?
Thanks,
Mariusz
On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg} wrote:
> Mariusz,
>
> You have uncovered another issue in MET related to precision. Your
data set certainly is testing our numerical
> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
> file and place the attached version in it's place. Then, run make
clean in the lib/vx_math folder, rebuild MET from the
> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
> square root error. Sorry for the trouble. Please let me know if
you have any questions.
>
> Paul
>
>
> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>
>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>> Queue: met_help
>> Subject: point stat question problem
>> Owner: Nobody
>> Requestors: Mariusz.Pagowski at noaa.gov
>> Status: new
>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>
>>
>> Paul et al,
>> I noticed that for QVAPOR and Q2 debiased square root error is zero
in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug or
something wrong with my processing.
>> Thanks,
>> Mariusz
>
>
>
> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
> // ** Copyright UCAR (c) 1992 - 2007
> // ** University Corporation for Atmospheric Research (UCAR)
> // ** National Center for Atmospheric Research (NCAR)
> // ** Research Applications Lab (RAL)
> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>
>
>
>
////////////////////////////////////////////////////////////////////////
>
>
> #ifndef __IS_BAD_DATA_H__
> #define __IS_BAD_DATA_H__
>
>
>
////////////////////////////////////////////////////////////////////////
>
>
> #include <cmath>
> #include "vx_math/constants.h"
>
>
>
////////////////////////////////////////////////////////////////////////
>
>
> inline int is_bad_data(int a) {
> if(a == bad_data_int) return(1);
> else return(0);
> }
>
> inline int is_bad_data(double a) {
> if(abs((long double) (a - bad_data_double)) < 10E-10) return(1);
> else return(0);
> }
>
> inline int is_bad_data(float a) {
> if(abs((long double) (a - bad_data_float)) < 10E-10) return(1);
> else return(0);
> }
>
> inline int is_eq(double a, double b, double tol) {
> if(abs((long double) (a - b)) < tol) return(1);
> else return(0);
> }
>
> inline int is_eq(double a, double b) {
> return(is_eq(a, b, 10E-10));
> }
>
>
////////////////////////////////////////////////////////////////////////
>
>
> #endif // __IS_BAD_DATA_H__
>
>
>
////////////////////////////////////////////////////////////////////////
>
>
>
>
------------------------------------------------
Subject: point stat question problem
From: Paul Oldenburg
Time: Fri Jun 17 07:46:15 2011
Mariusz,
I tared, zipped and attached the file.
Paul
On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>
> Paul,
> the same problem with Mac e-mail. Can you send the file separately,
not inside the text?
> Thanks,
> Mariusz
> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>
>> Mariusz,
>>
>> You have uncovered another issue in MET related to precision. Your
data set certainly is testing our numerical
>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>> file and place the attached version in it's place. Then, run make
clean in the lib/vx_math folder, rebuild MET from the
>> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
>> square root error. Sorry for the trouble. Please let me know if
you have any questions.
>>
>> Paul
>>
>>
>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>
>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>> Queue: met_help
>>> Subject: point stat question problem
>>> Owner: Nobody
>>> Requestors: Mariusz.Pagowski at noaa.gov
>>> Status: new
>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>
>>>
>>> Paul et al,
>>> I noticed that for QVAPOR and Q2 debiased square root error is
zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug
or something wrong with my processing.
>>> Thanks,
>>> Mariusz
>>
>>
>>
>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>> // ** Copyright UCAR (c) 1992 - 2007
>> // ** University Corporation for Atmospheric Research (UCAR)
>> // ** National Center for Atmospheric Research (NCAR)
>> // ** Research Applications Lab (RAL)
>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>
>>
>>
>>
////////////////////////////////////////////////////////////////////////
>>
>>
>> #ifndef __IS_BAD_DATA_H__
>> #define __IS_BAD_DATA_H__
>>
>>
>>
////////////////////////////////////////////////////////////////////////
>>
>>
>> #include <cmath>
>> #include "vx_math/constants.h"
>>
>>
>>
////////////////////////////////////////////////////////////////////////
>>
>>
>> inline int is_bad_data(int a) {
>> if(a == bad_data_int) return(1);
>> else return(0);
>> }
>>
>> inline int is_bad_data(double a) {
>> if(abs((long double) (a - bad_data_double)) < 10E-10) return(1);
>> else return(0);
>> }
>>
>> inline int is_bad_data(float a) {
>> if(abs((long double) (a - bad_data_float)) < 10E-10) return(1);
>> else return(0);
>> }
>>
>> inline int is_eq(double a, double b, double tol) {
>> if(abs((long double) (a - b)) < tol) return(1);
>> else return(0);
>> }
>>
>> inline int is_eq(double a, double b) {
>> return(is_eq(a, b, 10E-10));
>> }
>>
>>
////////////////////////////////////////////////////////////////////////
>>
>>
>> #endif // __IS_BAD_DATA_H__
>>
>>
>>
////////////////////////////////////////////////////////////////////////
>>
>>
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Mariusz Pagowski
Time: Fri Jun 17 09:42:13 2011
Thanks Paul,
I recompiled math library and the top of the code. There is one
improvement in running point_stat: P. correlations for QVAPOR for
500mb where
-99999.9 - now they are good numbers. But BCMSE for both QVAPOR and Q2
is still 0.00000
Mariusz
On Jun 17, 2011, at 7:46 AM, RAL HelpDesk {for Paul Oldenburg} wrote:
> Mariusz,
>
> I tared, zipped and attached the file.
>
> Paul
>
>
> On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>
>> Paul,
>> the same problem with Mac e-mail. Can you send the file separately,
not inside the text?
>> Thanks,
>> Mariusz
>> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>
>>> Mariusz,
>>>
>>> You have uncovered another issue in MET related to precision.
Your data set certainly is testing our numerical
>>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>>> file and place the attached version in it's place. Then, run make
clean in the lib/vx_math folder, rebuild MET from the
>>> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
>>> square root error. Sorry for the trouble. Please let me know if
you have any questions.
>>>
>>> Paul
>>>
>>>
>>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>>
>>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>> Queue: met_help
>>>> Subject: point stat question problem
>>>> Owner: Nobody
>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>> Status: new
>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>
>>>>
>>>> Paul et al,
>>>> I noticed that for QVAPOR and Q2 debiased square root error is
zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug
or something wrong with my processing.
>>>> Thanks,
>>>> Mariusz
>>>
>>>
>>>
>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>> // ** Copyright UCAR (c) 1992 - 2007
>>> // ** University Corporation for Atmospheric Research (UCAR)
>>> // ** National Center for Atmospheric Research (NCAR)
>>> // ** Research Applications Lab (RAL)
>>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>
>>>
>>>
>>>
////////////////////////////////////////////////////////////////////////
>>>
>>>
>>> #ifndef __IS_BAD_DATA_H__
>>> #define __IS_BAD_DATA_H__
>>>
>>>
>>>
////////////////////////////////////////////////////////////////////////
>>>
>>>
>>> #include <cmath>
>>> #include "vx_math/constants.h"
>>>
>>>
>>>
////////////////////////////////////////////////////////////////////////
>>>
>>>
>>> inline int is_bad_data(int a) {
>>> if(a == bad_data_int) return(1);
>>> else return(0);
>>> }
>>>
>>> inline int is_bad_data(double a) {
>>> if(abs((long double) (a - bad_data_double)) < 10E-10) return(1);
>>> else return(0);
>>> }
>>>
>>> inline int is_bad_data(float a) {
>>> if(abs((long double) (a - bad_data_float)) < 10E-10) return(1);
>>> else return(0);
>>> }
>>>
>>> inline int is_eq(double a, double b, double tol) {
>>> if(abs((long double) (a - b)) < tol) return(1);
>>> else return(0);
>>> }
>>>
>>> inline int is_eq(double a, double b) {
>>> return(is_eq(a, b, 10E-10));
>>> }
>>>
>>>
////////////////////////////////////////////////////////////////////////
>>>
>>>
>>> #endif // __IS_BAD_DATA_H__
>>>
>>>
>>>
////////////////////////////////////////////////////////////////////////
>>>
>>>
>>>
>>>
>>
>
>
> <is_bad_data.tar.gz>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Paul Oldenburg
Time: Fri Jun 17 09:44:52 2011
Mariusz,
Can you please send us the data, config file and command that you are
using to generate the data of interest? If the
data is too large to send via email, please follow the instructions at
the following location to upload it to our website:
http://www.dtcenter.org/met/users/support/met_help.php#ftp
Thanks,
Paul
On 06/17/2011 09:42 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>
> Thanks Paul,
> I recompiled math library and the top of the code. There is one
improvement in running point_stat: P. correlations for QVAPOR for
500mb where
> -99999.9 - now they are good numbers. But BCMSE for both QVAPOR and
Q2 is still 0.00000
> Mariusz
>
>
> On Jun 17, 2011, at 7:46 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>
>> Mariusz,
>>
>> I tared, zipped and attached the file.
>>
>> Paul
>>
>>
>> On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>
>>> Paul,
>>> the same problem with Mac e-mail. Can you send the file
separately, not inside the text?
>>> Thanks,
>>> Mariusz
>>> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>
>>>> Mariusz,
>>>>
>>>> You have uncovered another issue in MET related to precision.
Your data set certainly is testing our numerical
>>>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>>>> file and place the attached version in it's place. Then, run
make clean in the lib/vx_math folder, rebuild MET from the
>>>> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
>>>> square root error. Sorry for the trouble. Please let me know if
you have any questions.
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>
>>>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>> Queue: met_help
>>>>> Subject: point stat question problem
>>>>> Owner: Nobody
>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>> Status: new
>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>
>>>>>
>>>>> Paul et al,
>>>>> I noticed that for QVAPOR and Q2 debiased square root error is
zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug
or something wrong with my processing.
>>>>> Thanks,
>>>>> Mariusz
>>>>
>>>>
>>>>
>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>> // ** Copyright UCAR (c) 1992 - 2007
>>>> // ** University Corporation for Atmospheric Research (UCAR)
>>>> // ** National Center for Atmospheric Research (NCAR)
>>>> // ** Research Applications Lab (RAL)
>>>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>
>>>>
>>>>
>>>>
////////////////////////////////////////////////////////////////////////
>>>>
>>>>
>>>> #ifndef __IS_BAD_DATA_H__
>>>> #define __IS_BAD_DATA_H__
>>>>
>>>>
>>>>
////////////////////////////////////////////////////////////////////////
>>>>
>>>>
>>>> #include <cmath>
>>>> #include "vx_math/constants.h"
>>>>
>>>>
>>>>
////////////////////////////////////////////////////////////////////////
>>>>
>>>>
>>>> inline int is_bad_data(int a) {
>>>> if(a == bad_data_int) return(1);
>>>> else return(0);
>>>> }
>>>>
>>>> inline int is_bad_data(double a) {
>>>> if(abs((long double) (a - bad_data_double)) < 10E-10) return(1);
>>>> else return(0);
>>>> }
>>>>
>>>> inline int is_bad_data(float a) {
>>>> if(abs((long double) (a - bad_data_float)) < 10E-10) return(1);
>>>> else return(0);
>>>> }
>>>>
>>>> inline int is_eq(double a, double b, double tol) {
>>>> if(abs((long double) (a - b)) < tol) return(1);
>>>> else return(0);
>>>> }
>>>>
>>>> inline int is_eq(double a, double b) {
>>>> return(is_eq(a, b, 10E-10));
>>>> }
>>>>
>>>>
////////////////////////////////////////////////////////////////////////
>>>>
>>>>
>>>> #endif // __IS_BAD_DATA_H__
>>>>
>>>>
>>>>
////////////////////////////////////////////////////////////////////////
>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>> <is_bad_data.tar.gz>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Mariusz Pagowski
Time: Fri Jun 17 10:21:10 2011
Paul,
I put my data and config on your ftp.
I use command
point_stat $wrffile $obsfile $configfile \
-outdir $STATSDIR -v 2
Mariusz
On Jun 17, 2011, at 9:44 AM, RAL HelpDesk {for Paul Oldenburg} wrote:
> Mariusz,
>
> Can you please send us the data, config file and command that you
are using to generate the data of interest? If the
> data is too large to send via email, please follow the instructions
at the following location to upload it to our website:
>
> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>
> Thanks,
>
> Paul
>
>
>
> On 06/17/2011 09:42 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>
>> Thanks Paul,
>> I recompiled math library and the top of the code. There is one
improvement in running point_stat: P. correlations for QVAPOR for
500mb where
>> -99999.9 - now they are good numbers. But BCMSE for both QVAPOR and
Q2 is still 0.00000
>> Mariusz
>>
>>
>> On Jun 17, 2011, at 7:46 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>
>>> Mariusz,
>>>
>>> I tared, zipped and attached the file.
>>>
>>> Paul
>>>
>>>
>>> On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>
>>>> Paul,
>>>> the same problem with Mac e-mail. Can you send the file
separately, not inside the text?
>>>> Thanks,
>>>> Mariusz
>>>> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>
>>>>> Mariusz,
>>>>>
>>>>> You have uncovered another issue in MET related to precision.
Your data set certainly is testing our numerical
>>>>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>>>>> file and place the attached version in it's place. Then, run
make clean in the lib/vx_math folder, rebuild MET from the
>>>>> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
>>>>> square root error. Sorry for the trouble. Please let me know
if you have any questions.
>>>>>
>>>>> Paul
>>>>>
>>>>>
>>>>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>
>>>>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>>> Queue: met_help
>>>>>> Subject: point stat question problem
>>>>>> Owner: Nobody
>>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>>> Status: new
>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>>
>>>>>>
>>>>>> Paul et al,
>>>>>> I noticed that for QVAPOR and Q2 debiased square root error is
zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug
or something wrong with my processing.
>>>>>> Thanks,
>>>>>> Mariusz
>>>>>
>>>>>
>>>>>
>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>> // ** Copyright UCAR (c) 1992 - 2007
>>>>> // ** University Corporation for Atmospheric Research (UCAR)
>>>>> // ** National Center for Atmospheric Research (NCAR)
>>>>> // ** Research Applications Lab (RAL)
>>>>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>
>>>>>
>>>>>
>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>
>>>>>
>>>>> #ifndef __IS_BAD_DATA_H__
>>>>> #define __IS_BAD_DATA_H__
>>>>>
>>>>>
>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>
>>>>>
>>>>> #include <cmath>
>>>>> #include "vx_math/constants.h"
>>>>>
>>>>>
>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>
>>>>>
>>>>> inline int is_bad_data(int a) {
>>>>> if(a == bad_data_int) return(1);
>>>>> else return(0);
>>>>> }
>>>>>
>>>>> inline int is_bad_data(double a) {
>>>>> if(abs((long double) (a - bad_data_double)) < 10E-10) return(1);
>>>>> else return(0);
>>>>> }
>>>>>
>>>>> inline int is_bad_data(float a) {
>>>>> if(abs((long double) (a - bad_data_float)) < 10E-10) return(1);
>>>>> else return(0);
>>>>> }
>>>>>
>>>>> inline int is_eq(double a, double b, double tol) {
>>>>> if(abs((long double) (a - b)) < tol) return(1);
>>>>> else return(0);
>>>>> }
>>>>>
>>>>> inline int is_eq(double a, double b) {
>>>>> return(is_eq(a, b, 10E-10));
>>>>> }
>>>>>
>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>
>>>>>
>>>>> #endif // __IS_BAD_DATA_H__
>>>>>
>>>>>
>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>>
>>> <is_bad_data.tar.gz>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Paul Oldenburg
Time: Fri Jun 17 10:37:04 2011
Mariusz,
When I download and try to extract the tar file that you placed on the
FTP site, I get the following error:
$ tar xzvf pagowski.tar.gz
tar: This does not look like a tar archive
tar: Skipping to next header
tar: Error exit delayed from previous errors
Do I remember correctly that you are running MET on jet? If so, can
you just tell me the paths and filenames of the
files you are using there? If you provide read access to all, I
should be able to grab them and study them.
Thanks,
Paul
On 06/17/2011 10:21 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>
> Paul,
> I put my data and config on your ftp.
> I use command
>
> point_stat $wrffile $obsfile $configfile \
> -outdir $STATSDIR -v 2
>
> Mariusz
>
> On Jun 17, 2011, at 9:44 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>
>> Mariusz,
>>
>> Can you please send us the data, config file and command that you
are using to generate the data of interest? If the
>> data is too large to send via email, please follow the instructions
at the following location to upload it to our website:
>>
>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>
>> Thanks,
>>
>> Paul
>>
>>
>>
>> On 06/17/2011 09:42 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>
>>> Thanks Paul,
>>> I recompiled math library and the top of the code. There is one
improvement in running point_stat: P. correlations for QVAPOR for
500mb where
>>> -99999.9 - now they are good numbers. But BCMSE for both QVAPOR
and Q2 is still 0.00000
>>> Mariusz
>>>
>>>
>>> On Jun 17, 2011, at 7:46 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>
>>>> Mariusz,
>>>>
>>>> I tared, zipped and attached the file.
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>
>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>
>>>>> Paul,
>>>>> the same problem with Mac e-mail. Can you send the file
separately, not inside the text?
>>>>> Thanks,
>>>>> Mariusz
>>>>> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>>
>>>>>> Mariusz,
>>>>>>
>>>>>> You have uncovered another issue in MET related to precision.
Your data set certainly is testing our numerical
>>>>>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>>>>>> file and place the attached version in it's place. Then, run
make clean in the lib/vx_math folder, rebuild MET from the
>>>>>> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
>>>>>> square root error. Sorry for the trouble. Please let me know
if you have any questions.
>>>>>>
>>>>>> Paul
>>>>>>
>>>>>>
>>>>>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>>
>>>>>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>>>> Queue: met_help
>>>>>>> Subject: point stat question problem
>>>>>>> Owner: Nobody
>>>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>>>> Status: new
>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>>>
>>>>>>>
>>>>>>> Paul et al,
>>>>>>> I noticed that for QVAPOR and Q2 debiased square root error is
zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a bug
or something wrong with my processing.
>>>>>>> Thanks,
>>>>>>> Mariusz
>>>>>>
>>>>>>
>>>>>>
>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>> // ** Copyright UCAR (c) 1992 - 2007
>>>>>> // ** University Corporation for Atmospheric Research (UCAR)
>>>>>> // ** National Center for Atmospheric Research (NCAR)
>>>>>> // ** Research Applications Lab (RAL)
>>>>>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>>
>>>>>>
>>>>>>
>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>
>>>>>>
>>>>>> #ifndef __IS_BAD_DATA_H__
>>>>>> #define __IS_BAD_DATA_H__
>>>>>>
>>>>>>
>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>
>>>>>>
>>>>>> #include <cmath>
>>>>>> #include "vx_math/constants.h"
>>>>>>
>>>>>>
>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>
>>>>>>
>>>>>> inline int is_bad_data(int a) {
>>>>>> if(a == bad_data_int) return(1);
>>>>>> else return(0);
>>>>>> }
>>>>>>
>>>>>> inline int is_bad_data(double a) {
>>>>>> if(abs((long double) (a - bad_data_double)) < 10E-10)
return(1);
>>>>>> else
return(0);
>>>>>> }
>>>>>>
>>>>>> inline int is_bad_data(float a) {
>>>>>> if(abs((long double) (a - bad_data_float)) < 10E-10) return(1);
>>>>>> else return(0);
>>>>>> }
>>>>>>
>>>>>> inline int is_eq(double a, double b, double tol) {
>>>>>> if(abs((long double) (a - b)) < tol) return(1);
>>>>>> else return(0);
>>>>>> }
>>>>>>
>>>>>> inline int is_eq(double a, double b) {
>>>>>> return(is_eq(a, b, 10E-10));
>>>>>> }
>>>>>>
>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>
>>>>>>
>>>>>> #endif // __IS_BAD_DATA_H__
>>>>>>
>>>>>>
>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>> <is_bad_data.tar.gz>
>>>
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Mariusz Pagowski
Time: Fri Jun 17 10:43:51 2011
Paul,
the path to my MET is
/lfs0/projects/chem-var/wrfvar/MET
files are in ./ftp
Thanks,
Mariusz
On Jun 17, 2011, at 10:37 AM, RAL HelpDesk {for Paul Oldenburg} wrote:
> Mariusz,
>
> When I download and try to extract the tar file that you placed on
the FTP site, I get the following error:
>
> $ tar xzvf pagowski.tar.gz
> tar: This does not look like a tar archive
> tar: Skipping to next header
> tar: Error exit delayed from previous errors
>
> Do I remember correctly that you are running MET on jet? If so, can
you just tell me the paths and filenames of the
> files you are using there? If you provide read access to all, I
should be able to grab them and study them.
>
> Thanks,
>
> Paul
>
>
> On 06/17/2011 10:21 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>
>> Paul,
>> I put my data and config on your ftp.
>> I use command
>>
>> point_stat $wrffile $obsfile $configfile \
>> -outdir $STATSDIR -v 2
>>
>> Mariusz
>>
>> On Jun 17, 2011, at 9:44 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>
>>> Mariusz,
>>>
>>> Can you please send us the data, config file and command that you
are using to generate the data of interest? If the
>>> data is too large to send via email, please follow the
instructions at the following location to upload it to our website:
>>>
>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>>
>>> Thanks,
>>>
>>> Paul
>>>
>>>
>>>
>>> On 06/17/2011 09:42 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>
>>>> Thanks Paul,
>>>> I recompiled math library and the top of the code. There is one
improvement in running point_stat: P. correlations for QVAPOR for
500mb where
>>>> -99999.9 - now they are good numbers. But BCMSE for both QVAPOR
and Q2 is still 0.00000
>>>> Mariusz
>>>>
>>>>
>>>> On Jun 17, 2011, at 7:46 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>
>>>>> Mariusz,
>>>>>
>>>>> I tared, zipped and attached the file.
>>>>>
>>>>> Paul
>>>>>
>>>>>
>>>>> On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>>
>>>>>> Paul,
>>>>>> the same problem with Mac e-mail. Can you send the file
separately, not inside the text?
>>>>>> Thanks,
>>>>>> Mariusz
>>>>>> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>>>
>>>>>>> Mariusz,
>>>>>>>
>>>>>>> You have uncovered another issue in MET related to precision.
Your data set certainly is testing our numerical
>>>>>>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>>>>>>> file and place the attached version in it's place. Then, run
make clean in the lib/vx_math folder, rebuild MET from the
>>>>>>> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
>>>>>>> square root error. Sorry for the trouble. Please let me know
if you have any questions.
>>>>>>>
>>>>>>> Paul
>>>>>>>
>>>>>>>
>>>>>>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>>>
>>>>>>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>>>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>>>>> Queue: met_help
>>>>>>>> Subject: point stat question problem
>>>>>>>> Owner: Nobody
>>>>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>>>>> Status: new
>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>>>>
>>>>>>>>
>>>>>>>> Paul et al,
>>>>>>>> I noticed that for QVAPOR and Q2 debiased square root error
is zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a
bug or something wrong with my processing.
>>>>>>>> Thanks,
>>>>>>>> Mariusz
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>>> // ** Copyright UCAR (c) 1992 - 2007
>>>>>>> // ** University Corporation for Atmospheric Research (UCAR)
>>>>>>> // ** National Center for Atmospheric Research (NCAR)
>>>>>>> // ** Research Applications Lab (RAL)
>>>>>>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>
>>>>>>>
>>>>>>> #ifndef __IS_BAD_DATA_H__
>>>>>>> #define __IS_BAD_DATA_H__
>>>>>>>
>>>>>>>
>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>
>>>>>>>
>>>>>>> #include <cmath>
>>>>>>> #include "vx_math/constants.h"
>>>>>>>
>>>>>>>
>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>
>>>>>>>
>>>>>>> inline int is_bad_data(int a) {
>>>>>>> if(a == bad_data_int) return(1);
>>>>>>> else return(0);
>>>>>>> }
>>>>>>>
>>>>>>> inline int is_bad_data(double a) {
>>>>>>> if(abs((long double) (a - bad_data_double)) < 10E-10)
return(1);
>>>>>>> else
return(0);
>>>>>>> }
>>>>>>>
>>>>>>> inline int is_bad_data(float a) {
>>>>>>> if(abs((long double) (a - bad_data_float)) < 10E-10)
return(1);
>>>>>>> else
return(0);
>>>>>>> }
>>>>>>>
>>>>>>> inline int is_eq(double a, double b, double tol) {
>>>>>>> if(abs((long double) (a - b)) < tol) return(1);
>>>>>>> else return(0);
>>>>>>> }
>>>>>>>
>>>>>>> inline int is_eq(double a, double b) {
>>>>>>> return(is_eq(a, b, 10E-10));
>>>>>>> }
>>>>>>>
>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>
>>>>>>>
>>>>>>> #endif // __IS_BAD_DATA_H__
>>>>>>>
>>>>>>>
>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> <is_bad_data.tar.gz>
>>>>
>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Paul Oldenburg
Time: Fri Jun 17 11:49:13 2011
Mariusz,
I stepped through your situation in the debugger and found that the
BCMSE is about 2e-6, which is smaller than the
precision used in the output file, which is 1e-5. If you want to
increase the precision of the scores reported in the
output files, change the following line of the source file
lib/vx_met_util/constants.h:
static const int default_precision = 5;
Set the 5 to whatever precision you want to view. Then, remove the
point_stat executable from the src/point_stat/
folder and rebuild MET. Please let me know if you have any questions.
Thanks,
Paul
On 06/17/2011 10:43 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>
> Paul,
> the path to my MET is
> /lfs0/projects/chem-var/wrfvar/MET
> files are in ./ftp
> Thanks,
> Mariusz
>
>
> On Jun 17, 2011, at 10:37 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>
>> Mariusz,
>>
>> When I download and try to extract the tar file that you placed on
the FTP site, I get the following error:
>>
>> $ tar xzvf pagowski.tar.gz
>> tar: This does not look like a tar archive
>> tar: Skipping to next header
>> tar: Error exit delayed from previous errors
>>
>> Do I remember correctly that you are running MET on jet? If so,
can you just tell me the paths and filenames of the
>> files you are using there? If you provide read access to all, I
should be able to grab them and study them.
>>
>> Thanks,
>>
>> Paul
>>
>>
>> On 06/17/2011 10:21 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>
>>> Paul,
>>> I put my data and config on your ftp.
>>> I use command
>>>
>>> point_stat $wrffile $obsfile $configfile \
>>> -outdir $STATSDIR -v 2
>>>
>>> Mariusz
>>>
>>> On Jun 17, 2011, at 9:44 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>
>>>> Mariusz,
>>>>
>>>> Can you please send us the data, config file and command that you
are using to generate the data of interest? If the
>>>> data is too large to send via email, please follow the
instructions at the following location to upload it to our website:
>>>>
>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>>>
>>>> Thanks,
>>>>
>>>> Paul
>>>>
>>>>
>>>>
>>>> On 06/17/2011 09:42 AM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>
>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>
>>>>> Thanks Paul,
>>>>> I recompiled math library and the top of the code. There is one
improvement in running point_stat: P. correlations for QVAPOR for
500mb where
>>>>> -99999.9 - now they are good numbers. But BCMSE for both QVAPOR
and Q2 is still 0.00000
>>>>> Mariusz
>>>>>
>>>>>
>>>>> On Jun 17, 2011, at 7:46 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>>
>>>>>> Mariusz,
>>>>>>
>>>>>> I tared, zipped and attached the file.
>>>>>>
>>>>>> Paul
>>>>>>
>>>>>>
>>>>>> On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>>
>>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600
>
>>>>>>>
>>>>>>> Paul,
>>>>>>> the same problem with Mac e-mail. Can you send the file
separately, not inside the text?
>>>>>>> Thanks,
>>>>>>> Mariusz
>>>>>>> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>>>>
>>>>>>>> Mariusz,
>>>>>>>>
>>>>>>>> You have uncovered another issue in MET related to precision.
Your data set certainly is testing our numerical
>>>>>>>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>>>>>>>> file and place the attached version in it's place. Then, run
make clean in the lib/vx_math folder, rebuild MET from the
>>>>>>>> top-level and rerun your QVAPOR and Q2 verification to see if
reasonable numbers appear in the output for debiased
>>>>>>>> square root error. Sorry for the trouble. Please let me
know if you have any questions.
>>>>>>>>
>>>>>>>> Paul
>>>>>>>>
>>>>>>>>
>>>>>>>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>>>>
>>>>>>>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>>>>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>>>>>> Queue: met_help
>>>>>>>>> Subject: point stat question problem
>>>>>>>>> Owner: Nobody
>>>>>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>>>>>> Status: new
>>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Paul et al,
>>>>>>>>> I noticed that for QVAPOR and Q2 debiased square root error
is zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a
bug or something wrong with my processing.
>>>>>>>>> Thanks,
>>>>>>>>> Mariusz
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>>>> // ** Copyright UCAR (c) 1992 - 2007
>>>>>>>> // ** University Corporation for Atmospheric Research (UCAR)
>>>>>>>> // ** National Center for Atmospheric Research (NCAR)
>>>>>>>> // ** Research Applications Lab (RAL)
>>>>>>>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>>>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>
>>>>>>>>
>>>>>>>> #ifndef __IS_BAD_DATA_H__
>>>>>>>> #define __IS_BAD_DATA_H__
>>>>>>>>
>>>>>>>>
>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>
>>>>>>>>
>>>>>>>> #include <cmath>
>>>>>>>> #include "vx_math/constants.h"
>>>>>>>>
>>>>>>>>
>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>
>>>>>>>>
>>>>>>>> inline int is_bad_data(int a) {
>>>>>>>> if(a == bad_data_int) return(1);
>>>>>>>> else return(0);
>>>>>>>> }
>>>>>>>>
>>>>>>>> inline int is_bad_data(double a) {
>>>>>>>> if(abs((long double) (a - bad_data_double)) < 10E-10)
return(1);
>>>>>>>> else
return(0);
>>>>>>>> }
>>>>>>>>
>>>>>>>> inline int is_bad_data(float a) {
>>>>>>>> if(abs((long double) (a - bad_data_float)) < 10E-10)
return(1);
>>>>>>>> else
return(0);
>>>>>>>> }
>>>>>>>>
>>>>>>>> inline int is_eq(double a, double b, double tol) {
>>>>>>>> if(abs((long double) (a - b)) < tol) return(1);
>>>>>>>> else return(0);
>>>>>>>> }
>>>>>>>>
>>>>>>>> inline int is_eq(double a, double b) {
>>>>>>>> return(is_eq(a, b, 10E-10));
>>>>>>>> }
>>>>>>>>
>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>
>>>>>>>>
>>>>>>>> #endif // __IS_BAD_DATA_H__
>>>>>>>>
>>>>>>>>
>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>> <is_bad_data.tar.gz>
>>>>>
>>>>
>>>>
>>>
>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #47600] point stat question problem
From: Mariusz Pagowski
Time: Fri Jun 17 11:57:10 2011
Paul,
thanks for solving that issue. It looks like BCMSE is a useless
quantity for mixing ratio in kg/kg because of the truncation issue.
Thanks again,
Mariusz
On Jun 17, 2011, at 11:49 AM, RAL HelpDesk {for Paul Oldenburg} wrote:
> Mariusz,
>
> I stepped through your situation in the debugger and found that the
BCMSE is about 2e-6, which is smaller than the
> precision used in the output file, which is 1e-5. If you want to
increase the precision of the scores reported in the
> output files, change the following line of the source file
lib/vx_met_util/constants.h:
>
> static const int default_precision = 5;
>
> Set the 5 to whatever precision you want to view. Then, remove the
point_stat executable from the src/point_stat/
> folder and rebuild MET. Please let me know if you have any
questions.
>
> Thanks,
>
> Paul
>
>
> On 06/17/2011 10:43 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>
>> Paul,
>> the path to my MET is
>> /lfs0/projects/chem-var/wrfvar/MET
>> files are in ./ftp
>> Thanks,
>> Mariusz
>>
>>
>> On Jun 17, 2011, at 10:37 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>
>>> Mariusz,
>>>
>>> When I download and try to extract the tar file that you placed on
the FTP site, I get the following error:
>>>
>>> $ tar xzvf pagowski.tar.gz
>>> tar: This does not look like a tar archive
>>> tar: Skipping to next header
>>> tar: Error exit delayed from previous errors
>>>
>>> Do I remember correctly that you are running MET on jet? If so,
can you just tell me the paths and filenames of the
>>> files you are using there? If you provide read access to all, I
should be able to grab them and study them.
>>>
>>> Thanks,
>>>
>>> Paul
>>>
>>>
>>> On 06/17/2011 10:21 AM, RAL HelpDesk {for Mariusz Pagowski} wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>
>>>> Paul,
>>>> I put my data and config on your ftp.
>>>> I use command
>>>>
>>>> point_stat $wrffile $obsfile $configfile \
>>>> -outdir $STATSDIR -v 2
>>>>
>>>> Mariusz
>>>>
>>>> On Jun 17, 2011, at 9:44 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>
>>>>> Mariusz,
>>>>>
>>>>> Can you please send us the data, config file and command that
you are using to generate the data of interest? If the
>>>>> data is too large to send via email, please follow the
instructions at the following location to upload it to our website:
>>>>>
>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Paul
>>>>>
>>>>>
>>>>>
>>>>> On 06/17/2011 09:42 AM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>>
>>>>>> Thanks Paul,
>>>>>> I recompiled math library and the top of the code. There is one
improvement in running point_stat: P. correlations for QVAPOR for
500mb where
>>>>>> -99999.9 - now they are good numbers. But BCMSE for both QVAPOR
and Q2 is still 0.00000
>>>>>> Mariusz
>>>>>>
>>>>>>
>>>>>> On Jun 17, 2011, at 7:46 AM, RAL HelpDesk {for Paul Oldenburg}
wrote:
>>>>>>
>>>>>>> Mariusz,
>>>>>>>
>>>>>>> I tared, zipped and attached the file.
>>>>>>>
>>>>>>> Paul
>>>>>>>
>>>>>>>
>>>>>>> On 06/16/2011 09:46 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>>>
>>>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600
>
>>>>>>>>
>>>>>>>> Paul,
>>>>>>>> the same problem with Mac e-mail. Can you send the file
separately, not inside the text?
>>>>>>>> Thanks,
>>>>>>>> Mariusz
>>>>>>>> On Jun 16, 2011, at 4:48 PM, RAL HelpDesk {for Paul
Oldenburg} wrote:
>>>>>>>>
>>>>>>>>> Mariusz,
>>>>>>>>>
>>>>>>>>> You have uncovered another issue in MET related to
precision. Your data set certainly is testing our numerical
>>>>>>>>> thresholds. I attached a copy of the MET source file
lib/vx_math/is_bad_data.h. Please move your current copy of this
>>>>>>>>> file and place the attached version in it's place. Then,
run make clean in the lib/vx_math folder, rebuild MET from the
>>>>>>>>> top-level and rerun your QVAPOR and Q2 verification to see
if reasonable numbers appear in the output for debiased
>>>>>>>>> square root error. Sorry for the trouble. Please let me
know if you have any questions.
>>>>>>>>>
>>>>>>>>> Paul
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 06/16/2011 02:35 PM, RAL HelpDesk {for Mariusz Pagowski}
wrote:
>>>>>>>>>>
>>>>>>>>>> Thu Jun 16 14:35:06 2011: Request 47600 was acted upon.
>>>>>>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>>>>>>> Queue: met_help
>>>>>>>>>> Subject: point stat question problem
>>>>>>>>>> Owner: Nobody
>>>>>>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>>>>>>> Status: new
>>>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=47600 >
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Paul et al,
>>>>>>>>>> I noticed that for QVAPOR and Q2 debiased square root error
is zero in my outputs (BCMSE, RMSE is ok). Can you check if it is a
bug or something wrong with my processing.
>>>>>>>>>> Thanks,
>>>>>>>>>> Mariusz
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>>>>> // ** Copyright UCAR (c) 1992 - 2007
>>>>>>>>> // ** University Corporation for Atmospheric Research (UCAR)
>>>>>>>>> // ** National Center for Atmospheric Research (NCAR)
>>>>>>>>> // ** Research Applications Lab (RAL)
>>>>>>>>> // ** P.O.Box 3000, Boulder, Colorado, 80307-3000, USA
>>>>>>>>> // *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> #ifndef __IS_BAD_DATA_H__
>>>>>>>>> #define __IS_BAD_DATA_H__
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> #include <cmath>
>>>>>>>>> #include "vx_math/constants.h"
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> inline int is_bad_data(int a) {
>>>>>>>>> if(a == bad_data_int) return(1);
>>>>>>>>> else return(0);
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> inline int is_bad_data(double a) {
>>>>>>>>> if(abs((long double) (a - bad_data_double)) < 10E-10)
return(1);
>>>>>>>>> else
return(0);
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> inline int is_bad_data(float a) {
>>>>>>>>> if(abs((long double) (a - bad_data_float)) < 10E-10)
return(1);
>>>>>>>>> else
return(0);
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> inline int is_eq(double a, double b, double tol) {
>>>>>>>>> if(abs((long double) (a - b)) < tol) return(1);
>>>>>>>>> else return(0);
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> inline int is_eq(double a, double b) {
>>>>>>>>> return(is_eq(a, b, 10E-10));
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> #endif // __IS_BAD_DATA_H__
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
////////////////////////////////////////////////////////////////////////
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> <is_bad_data.tar.gz>
>>>>>>
>>>>>
>>>>>
>>>>
>>>
>>>
>>
>>
>
>
------------------------------------------------
More information about the Met_help
mailing list