[Met_help] [rt.rap.ucar.edu #52019] History for use point_stat for multiple forecasts
Paul Oldenburg via RT
met_help at ucar.edu
Thu Dec 29 08:36:57 MST 2011
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
>
>
> Hi,
>
> We are interested in using point_stat for the multiple forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>
> Is this a correct command?
>
> point_stat forecast_dir observation_dir PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end 2010529_18
>
> How about using –fcst_valid and –fcst_lead time?
>
> In addition, is the output of this analysis for each time stamp or for the whole time?
>
>
> Thanks,
>
>
> Shujia
>
>
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Wed Dec 14 10:23:04 2011
Shujia,
The point_stat tool was designed to perform comparisons for a single
valid time. The best approach in your situation
would be to call point_stat once for each forecast valid time that you
list. The obs_valid_beg/end options allow the
user to control the window of observation valid time that will match
the forecast valid time. The fcst_valid/lead
options allow the user to specify which forecast valid time to verify
in an input file that contains multiple valid
times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
if you have any other questions.
Thanks,
Paul
On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>
> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
> Queue: met_help
> Subject: use point_stat for multiple forecasts
> Owner: Nobody
> Requestors: shujia.zhou-1 at nasa.gov
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>
>
>>
>>
>> Hi,
>>
>> We are interested in using point_stat for the multiple forecasts.
For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>
>> Is this a correct command?
>>
>> point_stat forecast_dir observation_dir PointStatConfig_SFC_WRF
–obs_valid_end 2010529_00 –obs_valid_end 2010529_18
>>
>> How about using –fcst_valid and –fcst_lead time?
>>
>> In addition, is the output of this analysis for each time stamp or
for the whole time?
>>
>>
>> Thanks,
>>
>>
>> Shujia
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Shujia Zhou
Time: Wed Dec 14 11:49:28 2011
Paul:
Thank you for your prompt reply.
I have two other questions:
1. Do you have a sample PointStatConfig file for surface wind and soil
moisture? Our specification does not work:
for wind:
fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
for soil:
fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)", "Q2(0,*,*)" ];
obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
in PB2NCConfig file, we used:
message_type[] = ["ANYSFC"];
obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
2. Is we want to have a particular region to compare with
observation, we just need use or create a
x.poly file for mask_poly=. Is that right?
Thanks,
Shujia
On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
> Shujia,
>
> The point_stat tool was designed to perform comparisons for a single
valid time. The best approach in your situation
> would be to call point_stat once for each forecast valid time that
you list. The obs_valid_beg/end options allow the
> user to control the window of observation valid time that will match
the forecast valid time. The fcst_valid/lead
> options allow the user to specify which forecast valid time to
verify in an input file that contains multiple valid
> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
> if you have any other questions.
>
> Thanks,
>
> Paul
>
>
> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>
>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>> Queue: met_help
>> Subject: use point_stat for multiple forecasts
>> Owner: Nobody
>> Requestors: shujia.zhou-1 at nasa.gov
>> Status: new
>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>
>>
>>>
>>>
>>> Hi,
>>>
>>> We are interested in using point_stat for the multiple forecasts.
For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>
>>> Is this a correct command?
>>>
>>> point_stat forecast_dir observation_dir PointStatConfig_SFC_WRF
–obs_valid_end 2010529_00 –obs_valid_end 2010529_18
>>>
>>> How about using –fcst_valid and –fcst_lead time?
>>>
>>> In addition, is the output of this analysis for each time stamp or
for the whole time?
>>>
>>>
>>> Thanks,
>>>
>>>
>>> Shujia
>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Wed Dec 14 13:00:06 2011
Shujia,
Can you please send me your model forecast data, obs data and config
file? I'll take a look at it and try to build a
config file for you to use. You can upload data to our FTP site using
these instructions:
http://www.dtcenter.org/met/users/support/met_help.php#ftp.
Thanks,
Paul
On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>
> Paul:
>
>
> Thank you for your prompt reply.
>
> I have two other questions:
>
> 1. Do you have a sample PointStatConfig file for surface wind and
soil moisture? Our specification does not work:
>
> for wind:
> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>
> for soil:
>
> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)", "Q2(0,*,*)" ];
>
> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>
> in PB2NCConfig file, we used:
>
> message_type[] = ["ANYSFC"];
>
> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>
>
>
> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
> x.poly file for mask_poly=. Is that right?
>
>
> Thanks,
>
>
> Shujia
>
>
> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>
>> Shujia,
>>
>> The point_stat tool was designed to perform comparisons for a
single valid time. The best approach in your situation
>> would be to call point_stat once for each forecast valid time that
you list. The obs_valid_beg/end options allow the
>> user to control the window of observation valid time that will
match the forecast valid time. The fcst_valid/lead
>> options allow the user to specify which forecast valid time to
verify in an input file that contains multiple valid
>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>> if you have any other questions.
>>
>> Thanks,
>>
>> Paul
>>
>>
>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>
>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>> Queue: met_help
>>> Subject: use point_stat for multiple forecasts
>>> Owner: Nobody
>>> Requestors: shujia.zhou-1 at nasa.gov
>>> Status: new
>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>
>>>
>>>>
>>>>
>>>> Hi,
>>>>
>>>> We are interested in using point_stat for the multiple forecasts.
For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>
>>>> Is this a correct command?
>>>>
>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>
>>>> How about using –fcst_valid and –fcst_lead time?
>>>>
>>>> In addition, is the output of this analysis for each time stamp
or for the whole time?
>>>>
>>>>
>>>> Thanks,
>>>>
>>>>
>>>> Shujia
>>>>
>>>>
>>>
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Shujia Zhou
Time: Wed Dec 14 14:53:54 2011
Paul:
I did put my data along with the scripts into /incoming/irap/met_help.
(Sorry I forgot putting it into zhou_data.). I would like to ask
you to see whether my script for T2 is correct or not although it is
working and the result is reasonable. For the soil_wind dir, I
put the script for soil and wind, they do not work.
Thanks,
Shujia
On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
> Shujia,
>
> Can you please send me your model forecast data, obs data and config
file? I'll take a look at it and try to build a
> config file for you to use. You can upload data to our FTP site
using these instructions:
> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>
> Thanks,
>
> Paul
>
>
> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>
>> Paul:
>>
>>
>> Thank you for your prompt reply.
>>
>> I have two other questions:
>>
>> 1. Do you have a sample PointStatConfig file for surface wind and
soil moisture? Our specification does not work:
>>
>> for wind:
>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>
>> for soil:
>>
>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)", "Q2(0,*,*)" ];
>>
>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>
>> in PB2NCConfig file, we used:
>>
>> message_type[] = ["ANYSFC"];
>>
>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>
>>
>>
>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>> x.poly file for mask_poly=. Is that right?
>>
>>
>> Thanks,
>>
>>
>> Shujia
>>
>>
>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>
>>> Shujia,
>>>
>>> The point_stat tool was designed to perform comparisons for a
single valid time. The best approach in your situation
>>> would be to call point_stat once for each forecast valid time that
you list. The obs_valid_beg/end options allow the
>>> user to control the window of observation valid time that will
match the forecast valid time. The fcst_valid/lead
>>> options allow the user to specify which forecast valid time to
verify in an input file that contains multiple valid
>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>> if you have any other questions.
>>>
>>> Thanks,
>>>
>>> Paul
>>>
>>>
>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>
>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>> Queue: met_help
>>>> Subject: use point_stat for multiple forecasts
>>>> Owner: Nobody
>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>> Status: new
>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>
>>>>
>>>>>
>>>>>
>>>>> Hi,
>>>>>
>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>
>>>>> Is this a correct command?
>>>>>
>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>
>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>
>>>>> In addition, is the output of this analysis for each time stamp
or for the whole time?
>>>>>
>>>>>
>>>>> Thanks,
>>>>>
>>>>>
>>>>> Shujia
>>>>>
>>>>>
>>>>
>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Shujia Zhou
Time: Wed Dec 14 15:17:44 2011
Paul:
The tarball is:
shujia_met_question.tar.gz
Shujia
On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
> Paul:
>
> I did put my data along with the scripts into
/incoming/irap/met_help.
> (Sorry I forgot putting it into zhou_data.). I would like to ask
> you to see whether my script for T2 is correct or not although it is
> working and the result is reasonable. For the soil_wind dir, I
> put the script for soil and wind, they do not work.
>
> Thanks,
>
>
> Shujia
> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>
>> Shujia,
>>
>> Can you please send me your model forecast data, obs data and
config file? I'll take a look at it and try to build a
>> config file for you to use. You can upload data to our FTP site
using these instructions:
>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>
>> Thanks,
>>
>> Paul
>>
>>
>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>
>>> Paul:
>>>
>>>
>>> Thank you for your prompt reply.
>>>
>>> I have two other questions:
>>>
>>> 1. Do you have a sample PointStatConfig file for surface wind and
soil moisture? Our specification does not work:
>>>
>>> for wind:
>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>
>>> for soil:
>>>
>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)", "Q2(0,*,*)" ];
>>>
>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>
>>> in PB2NCConfig file, we used:
>>>
>>> message_type[] = ["ANYSFC"];
>>>
>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>>
>>>
>>>
>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>> x.poly file for mask_poly=. Is that right?
>>>
>>>
>>> Thanks,
>>>
>>>
>>> Shujia
>>>
>>>
>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>
>>>> Shujia,
>>>>
>>>> The point_stat tool was designed to perform comparisons for a
single valid time. The best approach in your situation
>>>> would be to call point_stat once for each forecast valid time
that you list. The obs_valid_beg/end options allow the
>>>> user to control the window of observation valid time that will
match the forecast valid time. The fcst_valid/lead
>>>> options allow the user to specify which forecast valid time to
verify in an input file that contains multiple valid
>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>> if you have any other questions.
>>>>
>>>> Thanks,
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>
>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>> Queue: met_help
>>>>> Subject: use point_stat for multiple forecasts
>>>>> Owner: Nobody
>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>> Status: new
>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>
>>>>>> Is this a correct command?
>>>>>>
>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>
>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>
>>>>>> In addition, is the output of this analysis for each time stamp
or for the whole time?
>>>>>>
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>>
>>>>>> Shujia
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Thu Dec 15 08:24:41 2011
Shujia,
The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
of WRF. MET cannot read such files, partly because they are on eta
levels. Do you have access to the WRF Unified Post
Processor (UPP)? If so, please post-process the WRF NetCDF files
using that tool. Then, MET will be able to read the
model data. If you have any questions about this, please let me know.
Thanks,
Paul
On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>
> Paul:
>
> The tarball is:
>
> shujia_met_question.tar.gz
>
>
> Shujia
> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>
>> Paul:
>>
>> I did put my data along with the scripts into
/incoming/irap/met_help.
>> (Sorry I forgot putting it into zhou_data.). I would like to ask
>> you to see whether my script for T2 is correct or not although it
is
>> working and the result is reasonable. For the soil_wind dir, I
>> put the script for soil and wind, they do not work.
>>
>> Thanks,
>>
>>
>> Shujia
>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>
>>> Shujia,
>>>
>>> Can you please send me your model forecast data, obs data and
config file? I'll take a look at it and try to build a
>>> config file for you to use. You can upload data to our FTP site
using these instructions:
>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>
>>> Thanks,
>>>
>>> Paul
>>>
>>>
>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>
>>>> Paul:
>>>>
>>>>
>>>> Thank you for your prompt reply.
>>>>
>>>> I have two other questions:
>>>>
>>>> 1. Do you have a sample PointStatConfig file for surface wind and
soil moisture? Our specification does not work:
>>>>
>>>> for wind:
>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>
>>>> for soil:
>>>>
>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)", "Q2(0,*,*)"
];
>>>>
>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>
>>>> in PB2NCConfig file, we used:
>>>>
>>>> message_type[] = ["ANYSFC"];
>>>>
>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>>>
>>>>
>>>>
>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>> x.poly file for mask_poly=. Is that right?
>>>>
>>>>
>>>> Thanks,
>>>>
>>>>
>>>> Shujia
>>>>
>>>>
>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>
>>>>> Shujia,
>>>>>
>>>>> The point_stat tool was designed to perform comparisons for a
single valid time. The best approach in your situation
>>>>> would be to call point_stat once for each forecast valid time
that you list. The obs_valid_beg/end options allow the
>>>>> user to control the window of observation valid time that will
match the forecast valid time. The fcst_valid/lead
>>>>> options allow the user to specify which forecast valid time to
verify in an input file that contains multiple valid
>>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>>> if you have any other questions.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Paul
>>>>>
>>>>>
>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>
>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>> Queue: met_help
>>>>>> Subject: use point_stat for multiple forecasts
>>>>>> Owner: Nobody
>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>> Status: new
>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>
>>>>>>> Is this a correct command?
>>>>>>>
>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>
>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>
>>>>>>> In addition, is the output of this analysis for each time
stamp or for the whole time?
>>>>>>>
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>>
>>>>>>> Shujia
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Shujia Zhou
Time: Thu Dec 15 08:48:47 2011
Paul:
I have the data file in pressure level into zhou_data. It is wrf-2010-
06-02-06-00-00-PLEV.
It should work.
Thanks,
Thanks,
Shujia
On Dec 15, 2011, at 10:24 AM, Paul Oldenburg via RT wrote:
> Shujia,
>
> The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
> of WRF. MET cannot read such files, partly because they are on eta
levels. Do you have access to the WRF Unified Post
> Processor (UPP)? If so, please post-process the WRF NetCDF files
using that tool. Then, MET will be able to read the
> model data. If you have any questions about this, please let me
know.
>
> Thanks,
>
> Paul
>
>
> On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>
>> Paul:
>>
>> The tarball is:
>>
>> shujia_met_question.tar.gz
>>
>>
>> Shujia
>> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>>
>>> Paul:
>>>
>>> I did put my data along with the scripts into
/incoming/irap/met_help.
>>> (Sorry I forgot putting it into zhou_data.). I would like to ask
>>> you to see whether my script for T2 is correct or not although it
is
>>> working and the result is reasonable. For the soil_wind dir, I
>>> put the script for soil and wind, they do not work.
>>>
>>> Thanks,
>>>
>>>
>>> Shujia
>>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>>
>>>> Shujia,
>>>>
>>>> Can you please send me your model forecast data, obs data and
config file? I'll take a look at it and try to build a
>>>> config file for you to use. You can upload data to our FTP site
using these instructions:
>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>>
>>>> Thanks,
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>>
>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>
>>>>> Paul:
>>>>>
>>>>>
>>>>> Thank you for your prompt reply.
>>>>>
>>>>> I have two other questions:
>>>>>
>>>>> 1. Do you have a sample PointStatConfig file for surface wind
and soil moisture? Our specification does not work:
>>>>>
>>>>> for wind:
>>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>>
>>>>> for soil:
>>>>>
>>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)", "Q2(0,*,*)"
];
>>>>>
>>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>>
>>>>> in PB2NCConfig file, we used:
>>>>>
>>>>> message_type[] = ["ANYSFC"];
>>>>>
>>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>>>>
>>>>>
>>>>>
>>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>>> x.poly file for mask_poly=. Is that right?
>>>>>
>>>>>
>>>>> Thanks,
>>>>>
>>>>>
>>>>> Shujia
>>>>>
>>>>>
>>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>>
>>>>>> Shujia,
>>>>>>
>>>>>> The point_stat tool was designed to perform comparisons for a
single valid time. The best approach in your situation
>>>>>> would be to call point_stat once for each forecast valid time
that you list. The obs_valid_beg/end options allow the
>>>>>> user to control the window of observation valid time that will
match the forecast valid time. The fcst_valid/lead
>>>>>> options allow the user to specify which forecast valid time to
verify in an input file that contains multiple valid
>>>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>>>> if you have any other questions.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> Paul
>>>>>>
>>>>>>
>>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>>
>>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>>> Queue: met_help
>>>>>>> Subject: use point_stat for multiple forecasts
>>>>>>> Owner: Nobody
>>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>>> Status: new
>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>
>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>>
>>>>>>>> Is this a correct command?
>>>>>>>>
>>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>>
>>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>>
>>>>>>>> In addition, is the output of this analysis for each time
stamp or for the whole time?
>>>>>>>>
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>>
>>>>>>>> Shujia
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Thu Dec 15 08:52:00 2011
Shujia,
The file wrf-2010-06-02-06-00-00-PLEV that you just sent also appears
to be the NetCDF output of WRF, which MET cannot
read. MET only supports the following types of model data: WPP/UPP
GRIB output and p_interp NetCDF output. We
recommend that users run UPP, since MET has the best support for GRIB.
Do you have access to either WPP/UPP or p_interp?
Paul
On 12/15/2011 08:48 AM, Shujia Zhou via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>
> Paul:
>
> I have the data file in pressure level into zhou_data. It is wrf-
2010-06-02-06-00-00-PLEV.
> It should work.
>
> Thanks,
>
>
> Thanks,
>
> Shujia
> On Dec 15, 2011, at 10:24 AM, Paul Oldenburg via RT wrote:
>
>> Shujia,
>>
>> The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
>> of WRF. MET cannot read such files, partly because they are on eta
levels. Do you have access to the WRF Unified Post
>> Processor (UPP)? If so, please post-process the WRF NetCDF files
using that tool. Then, MET will be able to read the
>> model data. If you have any questions about this, please let me
know.
>>
>> Thanks,
>>
>> Paul
>>
>>
>> On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>
>>> Paul:
>>>
>>> The tarball is:
>>>
>>> shujia_met_question.tar.gz
>>>
>>>
>>> Shujia
>>> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>>>
>>>> Paul:
>>>>
>>>> I did put my data along with the scripts into
/incoming/irap/met_help.
>>>> (Sorry I forgot putting it into zhou_data.). I would like to ask
>>>> you to see whether my script for T2 is correct or not although it
is
>>>> working and the result is reasonable. For the soil_wind dir, I
>>>> put the script for soil and wind, they do not work.
>>>>
>>>> Thanks,
>>>>
>>>>
>>>> Shujia
>>>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>>>
>>>>> Shujia,
>>>>>
>>>>> Can you please send me your model forecast data, obs data and
config file? I'll take a look at it and try to build a
>>>>> config file for you to use. You can upload data to our FTP site
using these instructions:
>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Paul
>>>>>
>>>>>
>>>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>
>>>>>> Paul:
>>>>>>
>>>>>>
>>>>>> Thank you for your prompt reply.
>>>>>>
>>>>>> I have two other questions:
>>>>>>
>>>>>> 1. Do you have a sample PointStatConfig file for surface wind
and soil moisture? Our specification does not work:
>>>>>>
>>>>>> for wind:
>>>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>>>
>>>>>> for soil:
>>>>>>
>>>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)", "Q2(0,*,*)"
];
>>>>>>
>>>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>>>
>>>>>> in PB2NCConfig file, we used:
>>>>>>
>>>>>> message_type[] = ["ANYSFC"];
>>>>>>
>>>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>>>>>
>>>>>>
>>>>>>
>>>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>>>> x.poly file for mask_poly=. Is that right?
>>>>>>
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>>
>>>>>> Shujia
>>>>>>
>>>>>>
>>>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>>>
>>>>>>> Shujia,
>>>>>>>
>>>>>>> The point_stat tool was designed to perform comparisons for a
single valid time. The best approach in your situation
>>>>>>> would be to call point_stat once for each forecast valid time
that you list. The obs_valid_beg/end options allow the
>>>>>>> user to control the window of observation valid time that will
match the forecast valid time. The fcst_valid/lead
>>>>>>> options allow the user to specify which forecast valid time to
verify in an input file that contains multiple valid
>>>>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>>>>> if you have any other questions.
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> Paul
>>>>>>>
>>>>>>>
>>>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>>>
>>>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>>>> Queue: met_help
>>>>>>>> Subject: use point_stat for multiple forecasts
>>>>>>>> Owner: Nobody
>>>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>>>> Status: new
>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>
>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>>>
>>>>>>>>> Is this a correct command?
>>>>>>>>>
>>>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>>>
>>>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>>>
>>>>>>>>> In addition, is the output of this analysis for each time
stamp or for the whole time?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Shujia
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Thu Dec 15 08:56:36 2011
Shujia,
I was just informed by my colleague that the file wrf-2010-06-02-06-
00-00-PLEV does appear to the output of p_interp. I
will analyze this file and respond when I have some results. Sorry
for the confusion.
Thanks,
Paul
On 12/15/2011 08:51 AM, Paul Oldenburg wrote:
> Shujia,
>
> The file wrf-2010-06-02-06-00-00-PLEV that you just sent also
appears to be the NetCDF output of WRF, which MET cannot
> read. MET only supports the following types of model data: WPP/UPP
GRIB output and p_interp NetCDF output. We
> recommend that users run UPP, since MET has the best support for
GRIB. Do you have access to either WPP/UPP or p_interp?
>
> Paul
>
>
> On 12/15/2011 08:48 AM, Shujia Zhou via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>
>> Paul:
>>
>> I have the data file in pressure level into zhou_data. It is wrf-
2010-06-02-06-00-00-PLEV.
>> It should work.
>>
>> Thanks,
>>
>>
>> Thanks,
>>
>> Shujia
>> On Dec 15, 2011, at 10:24 AM, Paul Oldenburg via RT wrote:
>>
>>> Shujia,
>>>
>>> The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
>>> of WRF. MET cannot read such files, partly because they are on
eta levels. Do you have access to the WRF Unified Post
>>> Processor (UPP)? If so, please post-process the WRF NetCDF files
using that tool. Then, MET will be able to read the
>>> model data. If you have any questions about this, please let me
know.
>>>
>>> Thanks,
>>>
>>> Paul
>>>
>>>
>>> On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>
>>>> Paul:
>>>>
>>>> The tarball is:
>>>>
>>>> shujia_met_question.tar.gz
>>>>
>>>>
>>>> Shujia
>>>> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>>>>
>>>>> Paul:
>>>>>
>>>>> I did put my data along with the scripts into
/incoming/irap/met_help.
>>>>> (Sorry I forgot putting it into zhou_data.). I would like to ask
>>>>> you to see whether my script for T2 is correct or not although
it is
>>>>> working and the result is reasonable. For the soil_wind dir, I
>>>>> put the script for soil and wind, they do not work.
>>>>>
>>>>> Thanks,
>>>>>
>>>>>
>>>>> Shujia
>>>>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>>>>
>>>>>> Shujia,
>>>>>>
>>>>>> Can you please send me your model forecast data, obs data and
config file? I'll take a look at it and try to build a
>>>>>> config file for you to use. You can upload data to our FTP
site using these instructions:
>>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> Paul
>>>>>>
>>>>>>
>>>>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>>>>
>>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019
>
>>>>>>>
>>>>>>> Paul:
>>>>>>>
>>>>>>>
>>>>>>> Thank you for your prompt reply.
>>>>>>>
>>>>>>> I have two other questions:
>>>>>>>
>>>>>>> 1. Do you have a sample PointStatConfig file for surface wind
and soil moisture? Our specification does not work:
>>>>>>>
>>>>>>> for wind:
>>>>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>>>>
>>>>>>> for soil:
>>>>>>>
>>>>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)",
"Q2(0,*,*)" ];
>>>>>>>
>>>>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>>>>
>>>>>>> in PB2NCConfig file, we used:
>>>>>>>
>>>>>>> message_type[] = ["ANYSFC"];
>>>>>>>
>>>>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>>>>> x.poly file for mask_poly=. Is that right?
>>>>>>>
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>>
>>>>>>> Shujia
>>>>>>>
>>>>>>>
>>>>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>>>>
>>>>>>>> Shujia,
>>>>>>>>
>>>>>>>> The point_stat tool was designed to perform comparisons for a
single valid time. The best approach in your situation
>>>>>>>> would be to call point_stat once for each forecast valid time
that you list. The obs_valid_beg/end options allow the
>>>>>>>> user to control the window of observation valid time that
will match the forecast valid time. The fcst_valid/lead
>>>>>>>> options allow the user to specify which forecast valid time
to verify in an input file that contains multiple valid
>>>>>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>>>>>> if you have any other questions.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>> Paul
>>>>>>>>
>>>>>>>>
>>>>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>>>>
>>>>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>>>>> Queue: met_help
>>>>>>>>> Subject: use point_stat for multiple forecasts
>>>>>>>>> Owner: Nobody
>>>>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>>>>> Status: new
>>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>>>>
>>>>>>>>>> Is this a correct command?
>>>>>>>>>>
>>>>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>>>>
>>>>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>>>>
>>>>>>>>>> In addition, is the output of this analysis for each time
stamp or for the whole time?
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Shujia
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Thu Dec 15 10:25:06 2011
Shujia,
I was able to successfully run point_stat for the model data and
prepbufr data that you sent. First, I had to make the
following change to your PB2NC config file:
// add SPFH
obs_grib_code[] = ["TMP", "WIND", "WDIR", "SPFH"];
I used the two commands that included below. Notice that I had to use
the -obs_valid_end flag in the point_stat command
because the PrepBUFR observations are valid at and around 20100602_00
and the model data is valid at 20100602_06.
$MET_BASE/bin/pb2nc \
prepbufr.gdas.20100602.t00z.nr \
prepbufr.gdas.20100602.t00z.nr.nc \
PB2NCConfig_G212_SFC_WRF_sjz \
-v 3
$MET_BASE/bin/point_stat \
wrf-2010-06-02-06-00-00-PLEV \
prepbufr.gdas.20100602.t00z.nr.nc \
PointStatConfig_SFC_WRF_sjz \
-obs_valid_beg 20100601_22 \
-obs_valid_end 20100602_02 \
-outdir . -v 3
On 12/15/2011 08:56 AM, Paul Oldenburg wrote:
> Shujia,
>
> I was just informed by my colleague that the file wrf-2010-06-02-06-
00-00-PLEV does appear to the output of p_interp. I
> will analyze this file and respond when I have some results. Sorry
for the confusion.
>
> Thanks,
>
> Paul
>
>
> On 12/15/2011 08:51 AM, Paul Oldenburg wrote:
>> Shujia,
>>
>> The file wrf-2010-06-02-06-00-00-PLEV that you just sent also
appears to be the NetCDF output of WRF, which MET cannot
>> read. MET only supports the following types of model data: WPP/UPP
GRIB output and p_interp NetCDF output. We
>> recommend that users run UPP, since MET has the best support for
GRIB. Do you have access to either WPP/UPP or p_interp?
>>
>> Paul
>>
>>
>> On 12/15/2011 08:48 AM, Shujia Zhou via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>
>>> Paul:
>>>
>>> I have the data file in pressure level into zhou_data. It is wrf-
2010-06-02-06-00-00-PLEV.
>>> It should work.
>>>
>>> Thanks,
>>>
>>>
>>> Thanks,
>>>
>>> Shujia
>>> On Dec 15, 2011, at 10:24 AM, Paul Oldenburg via RT wrote:
>>>
>>>> Shujia,
>>>>
>>>> The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
>>>> of WRF. MET cannot read such files, partly because they are on
eta levels. Do you have access to the WRF Unified Post
>>>> Processor (UPP)? If so, please post-process the WRF NetCDF files
using that tool. Then, MET will be able to read the
>>>> model data. If you have any questions about this, please let me
know.
>>>>
>>>> Thanks,
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>>>>>
>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>
>>>>> Paul:
>>>>>
>>>>> The tarball is:
>>>>>
>>>>> shujia_met_question.tar.gz
>>>>>
>>>>>
>>>>> Shujia
>>>>> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>>>>>
>>>>>> Paul:
>>>>>>
>>>>>> I did put my data along with the scripts into
/incoming/irap/met_help.
>>>>>> (Sorry I forgot putting it into zhou_data.). I would like to
ask
>>>>>> you to see whether my script for T2 is correct or not although
it is
>>>>>> working and the result is reasonable. For the soil_wind dir, I
>>>>>> put the script for soil and wind, they do not work.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>>
>>>>>> Shujia
>>>>>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>>>>>
>>>>>>> Shujia,
>>>>>>>
>>>>>>> Can you please send me your model forecast data, obs data and
config file? I'll take a look at it and try to build a
>>>>>>> config file for you to use. You can upload data to our FTP
site using these instructions:
>>>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> Paul
>>>>>>>
>>>>>>>
>>>>>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>>>>>
>>>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019
>
>>>>>>>>
>>>>>>>> Paul:
>>>>>>>>
>>>>>>>>
>>>>>>>> Thank you for your prompt reply.
>>>>>>>>
>>>>>>>> I have two other questions:
>>>>>>>>
>>>>>>>> 1. Do you have a sample PointStatConfig file for surface wind
and soil moisture? Our specification does not work:
>>>>>>>>
>>>>>>>> for wind:
>>>>>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>>>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>>>>>
>>>>>>>> for soil:
>>>>>>>>
>>>>>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)",
"Q2(0,*,*)" ];
>>>>>>>>
>>>>>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>>>>>
>>>>>>>> in PB2NCConfig file, we used:
>>>>>>>>
>>>>>>>> message_type[] = ["ANYSFC"];
>>>>>>>>
>>>>>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>>>>>> x.poly file for mask_poly=. Is that right?
>>>>>>>>
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>>
>>>>>>>> Shujia
>>>>>>>>
>>>>>>>>
>>>>>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>>>>>
>>>>>>>>> Shujia,
>>>>>>>>>
>>>>>>>>> The point_stat tool was designed to perform comparisons for
a single valid time. The best approach in your situation
>>>>>>>>> would be to call point_stat once for each forecast valid
time that you list. The obs_valid_beg/end options allow the
>>>>>>>>> user to control the window of observation valid time that
will match the forecast valid time. The fcst_valid/lead
>>>>>>>>> options allow the user to specify which forecast valid time
to verify in an input file that contains multiple valid
>>>>>>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>>>>>>> if you have any other questions.
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>>
>>>>>>>>> Paul
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>>>>>
>>>>>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>>>>>> Queue: met_help
>>>>>>>>>> Subject: use point_stat for multiple forecasts
>>>>>>>>>> Owner: Nobody
>>>>>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>>>>>> Status: new
>>>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Hi,
>>>>>>>>>>>
>>>>>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>>>>>
>>>>>>>>>>> Is this a correct command?
>>>>>>>>>>>
>>>>>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>>>>>
>>>>>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>>>>>
>>>>>>>>>>> In addition, is the output of this analysis for each time
stamp or for the whole time?
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Shujia
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Thu Dec 15 10:28:18 2011
Shujia,
I'm sorry for sending this last email prematurely. I meant to add
that soil temperature and moisture observations
cannot be extracted or derived from PrepBUFR observations processed
with pb2nc. You will have to find another source of
observations if you want to verify those quantities. Please let me
know if you have any other questions.
Thanks,
Paul
On 12/15/2011 10:25 AM, Paul Oldenburg wrote:
> Shujia,
>
> I was able to successfully run point_stat for the model data and
prepbufr data that you sent. First, I had to make the
> following change to your PB2NC config file:
>
> // add SPFH
> obs_grib_code[] = ["TMP", "WIND", "WDIR", "SPFH"];
>
>
> I used the two commands that included below. Notice that I had to
use the -obs_valid_end flag in the point_stat command
> because the PrepBUFR observations are valid at and around
20100602_00 and the model data is valid at 20100602_06.
>
>
>
> $MET_BASE/bin/pb2nc \
> prepbufr.gdas.20100602.t00z.nr \
> prepbufr.gdas.20100602.t00z.nr.nc \
> PB2NCConfig_G212_SFC_WRF_sjz \
> -v 3
>
> $MET_BASE/bin/point_stat \
> wrf-2010-06-02-06-00-00-PLEV \
> prepbufr.gdas.20100602.t00z.nr.nc \
> PointStatConfig_SFC_WRF_sjz \
> -obs_valid_beg 20100601_22 \
> -obs_valid_end 20100602_02 \
> -outdir . -v 3
>
> On 12/15/2011 08:56 AM, Paul Oldenburg wrote:
>> Shujia,
>>
>> I was just informed by my colleague that the file wrf-2010-06-02-
06-00-00-PLEV does appear to the output of p_interp. I
>> will analyze this file and respond when I have some results. Sorry
for the confusion.
>>
>> Thanks,
>>
>> Paul
>>
>>
>> On 12/15/2011 08:51 AM, Paul Oldenburg wrote:
>>> Shujia,
>>>
>>> The file wrf-2010-06-02-06-00-00-PLEV that you just sent also
appears to be the NetCDF output of WRF, which MET cannot
>>> read. MET only supports the following types of model data:
WPP/UPP GRIB output and p_interp NetCDF output. We
>>> recommend that users run UPP, since MET has the best support for
GRIB. Do you have access to either WPP/UPP or p_interp?
>>>
>>> Paul
>>>
>>>
>>> On 12/15/2011 08:48 AM, Shujia Zhou via RT wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>
>>>> Paul:
>>>>
>>>> I have the data file in pressure level into zhou_data. It is wrf-
2010-06-02-06-00-00-PLEV.
>>>> It should work.
>>>>
>>>> Thanks,
>>>>
>>>>
>>>> Thanks,
>>>>
>>>> Shujia
>>>> On Dec 15, 2011, at 10:24 AM, Paul Oldenburg via RT wrote:
>>>>
>>>>> Shujia,
>>>>>
>>>>> The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
>>>>> of WRF. MET cannot read such files, partly because they are on
eta levels. Do you have access to the WRF Unified Post
>>>>> Processor (UPP)? If so, please post-process the WRF NetCDF
files using that tool. Then, MET will be able to read the
>>>>> model data. If you have any questions about this, please let me
know.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Paul
>>>>>
>>>>>
>>>>> On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>
>>>>>> Paul:
>>>>>>
>>>>>> The tarball is:
>>>>>>
>>>>>> shujia_met_question.tar.gz
>>>>>>
>>>>>>
>>>>>> Shujia
>>>>>> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>>>>>>
>>>>>>> Paul:
>>>>>>>
>>>>>>> I did put my data along with the scripts into
/incoming/irap/met_help.
>>>>>>> (Sorry I forgot putting it into zhou_data.). I would like to
ask
>>>>>>> you to see whether my script for T2 is correct or not although
it is
>>>>>>> working and the result is reasonable. For the soil_wind dir, I
>>>>>>> put the script for soil and wind, they do not work.
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>>
>>>>>>> Shujia
>>>>>>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>>>>>>
>>>>>>>> Shujia,
>>>>>>>>
>>>>>>>> Can you please send me your model forecast data, obs data and
config file? I'll take a look at it and try to build a
>>>>>>>> config file for you to use. You can upload data to our FTP
site using these instructions:
>>>>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>> Paul
>>>>>>>>
>>>>>>>>
>>>>>>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>>>>>>
>>>>>>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>
>>>>>>>>> Paul:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Thank you for your prompt reply.
>>>>>>>>>
>>>>>>>>> I have two other questions:
>>>>>>>>>
>>>>>>>>> 1. Do you have a sample PointStatConfig file for surface
wind and soil moisture? Our specification does not work:
>>>>>>>>>
>>>>>>>>> for wind:
>>>>>>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>>>>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>>>>>>
>>>>>>>>> for soil:
>>>>>>>>>
>>>>>>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)",
"Q2(0,*,*)" ];
>>>>>>>>>
>>>>>>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>>>>>>
>>>>>>>>> in PB2NCConfig file, we used:
>>>>>>>>>
>>>>>>>>> message_type[] = ["ANYSFC"];
>>>>>>>>>
>>>>>>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL", "SOILW"];
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>>>>>>> x.poly file for mask_poly=. Is that right?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Shujia
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>>>>>>
>>>>>>>>>> Shujia,
>>>>>>>>>>
>>>>>>>>>> The point_stat tool was designed to perform comparisons for
a single valid time. The best approach in your situation
>>>>>>>>>> would be to call point_stat once for each forecast valid
time that you list. The obs_valid_beg/end options allow the
>>>>>>>>>> user to control the window of observation valid time that
will match the forecast valid time. The fcst_valid/lead
>>>>>>>>>> options allow the user to specify which forecast valid time
to verify in an input file that contains multiple valid
>>>>>>>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>>>>>>>> if you have any other questions.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>>
>>>>>>>>>> Paul
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>>>>>>
>>>>>>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>>>>>>> Queue: met_help
>>>>>>>>>>> Subject: use point_stat for multiple forecasts
>>>>>>>>>>> Owner: Nobody
>>>>>>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>>>>>>> Status: new
>>>>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Hi,
>>>>>>>>>>>>
>>>>>>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>>>>>>
>>>>>>>>>>>> Is this a correct command?
>>>>>>>>>>>>
>>>>>>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>>>>>>
>>>>>>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>>>>>>
>>>>>>>>>>>> In addition, is the output of this analysis for each time
stamp or for the whole time?
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Shujia
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Shujia Zhou
Time: Thu Dec 15 15:00:09 2011
Paul:
Have you verified my scripts for T2? Is that correct?
Thanks,
Shujia
On Dec 15, 2011, at 12:28 PM, Paul Oldenburg via RT wrote:
> Shujia,
>
> I'm sorry for sending this last email prematurely. I meant to add
that soil temperature and moisture observations
> cannot be extracted or derived from PrepBUFR observations processed
with pb2nc. You will have to find another source of
> observations if you want to verify those quantities. Please let me
know if you have any other questions.
>
> Thanks,
>
> Paul
>
>
> On 12/15/2011 10:25 AM, Paul Oldenburg wrote:
>> Shujia,
>>
>> I was able to successfully run point_stat for the model data and
prepbufr data that you sent. First, I had to make the
>> following change to your PB2NC config file:
>>
>> // add SPFH
>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "SPFH"];
>>
>>
>> I used the two commands that included below. Notice that I had to
use the -obs_valid_end flag in the point_stat command
>> because the PrepBUFR observations are valid at and around
20100602_00 and the model data is valid at 20100602_06.
>>
>>
>>
>> $MET_BASE/bin/pb2nc \
>> prepbufr.gdas.20100602.t00z.nr \
>> prepbufr.gdas.20100602.t00z.nr.nc \
>> PB2NCConfig_G212_SFC_WRF_sjz \
>> -v 3
>>
>> $MET_BASE/bin/point_stat \
>> wrf-2010-06-02-06-00-00-PLEV \
>> prepbufr.gdas.20100602.t00z.nr.nc \
>> PointStatConfig_SFC_WRF_sjz \
>> -obs_valid_beg 20100601_22 \
>> -obs_valid_end 20100602_02 \
>> -outdir . -v 3
>>
>> On 12/15/2011 08:56 AM, Paul Oldenburg wrote:
>>> Shujia,
>>>
>>> I was just informed by my colleague that the file wrf-2010-06-02-
06-00-00-PLEV does appear to the output of p_interp. I
>>> will analyze this file and respond when I have some results.
Sorry for the confusion.
>>>
>>> Thanks,
>>>
>>> Paul
>>>
>>>
>>> On 12/15/2011 08:51 AM, Paul Oldenburg wrote:
>>>> Shujia,
>>>>
>>>> The file wrf-2010-06-02-06-00-00-PLEV that you just sent also
appears to be the NetCDF output of WRF, which MET cannot
>>>> read. MET only supports the following types of model data:
WPP/UPP GRIB output and p_interp NetCDF output. We
>>>> recommend that users run UPP, since MET has the best support for
GRIB. Do you have access to either WPP/UPP or p_interp?
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 12/15/2011 08:48 AM, Shujia Zhou via RT wrote:
>>>>>
>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>
>>>>> Paul:
>>>>>
>>>>> I have the data file in pressure level into zhou_data. It is
wrf-2010-06-02-06-00-00-PLEV.
>>>>> It should work.
>>>>>
>>>>> Thanks,
>>>>>
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Shujia
>>>>> On Dec 15, 2011, at 10:24 AM, Paul Oldenburg via RT wrote:
>>>>>
>>>>>> Shujia,
>>>>>>
>>>>>> The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
>>>>>> of WRF. MET cannot read such files, partly because they are on
eta levels. Do you have access to the WRF Unified Post
>>>>>> Processor (UPP)? If so, please post-process the WRF NetCDF
files using that tool. Then, MET will be able to read the
>>>>>> model data. If you have any questions about this, please let
me know.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> Paul
>>>>>>
>>>>>>
>>>>>> On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>>>>>>>
>>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019
>
>>>>>>>
>>>>>>> Paul:
>>>>>>>
>>>>>>> The tarball is:
>>>>>>>
>>>>>>> shujia_met_question.tar.gz
>>>>>>>
>>>>>>>
>>>>>>> Shujia
>>>>>>> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>>>>>>>
>>>>>>>> Paul:
>>>>>>>>
>>>>>>>> I did put my data along with the scripts into
/incoming/irap/met_help.
>>>>>>>> (Sorry I forgot putting it into zhou_data.). I would like to
ask
>>>>>>>> you to see whether my script for T2 is correct or not
although it is
>>>>>>>> working and the result is reasonable. For the soil_wind dir,
I
>>>>>>>> put the script for soil and wind, they do not work.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>>
>>>>>>>> Shujia
>>>>>>>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>>>>>>>
>>>>>>>>> Shujia,
>>>>>>>>>
>>>>>>>>> Can you please send me your model forecast data, obs data
and config file? I'll take a look at it and try to build a
>>>>>>>>> config file for you to use. You can upload data to our FTP
site using these instructions:
>>>>>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>>
>>>>>>>>> Paul
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>>>>>>>
>>>>>>>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>>
>>>>>>>>>> Paul:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Thank you for your prompt reply.
>>>>>>>>>>
>>>>>>>>>> I have two other questions:
>>>>>>>>>>
>>>>>>>>>> 1. Do you have a sample PointStatConfig file for surface
wind and soil moisture? Our specification does not work:
>>>>>>>>>>
>>>>>>>>>> for wind:
>>>>>>>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>>>>>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>>>>>>>
>>>>>>>>>> for soil:
>>>>>>>>>>
>>>>>>>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)",
"Q2(0,*,*)" ];
>>>>>>>>>>
>>>>>>>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>>>>>>>
>>>>>>>>>> in PB2NCConfig file, we used:
>>>>>>>>>>
>>>>>>>>>> message_type[] = ["ANYSFC"];
>>>>>>>>>>
>>>>>>>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL",
"SOILW"];
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>>>>>>>> x.poly file for mask_poly=. Is that right?
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Shujia
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>>>>>>>
>>>>>>>>>>> Shujia,
>>>>>>>>>>>
>>>>>>>>>>> The point_stat tool was designed to perform comparisons
for a single valid time. The best approach in your situation
>>>>>>>>>>> would be to call point_stat once for each forecast valid
time that you list. The obs_valid_beg/end options allow the
>>>>>>>>>>> user to control the window of observation valid time that
will match the forecast valid time. The fcst_valid/lead
>>>>>>>>>>> options allow the user to specify which forecast valid
time to verify in an input file that contains multiple valid
>>>>>>>>>>> times (usually in the form of many lead times for a single
initialization). Does that make sense? Please let me know
>>>>>>>>>>> if you have any other questions.
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>>
>>>>>>>>>>> Paul
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>>>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>>>>>>>> Queue: met_help
>>>>>>>>>>>> Subject: use point_stat for multiple forecasts
>>>>>>>>>>>> Owner: Nobody
>>>>>>>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>>>>>>>> Status: new
>>>>>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Hi,
>>>>>>>>>>>>>
>>>>>>>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Is this a correct command?
>>>>>>>>>>>>>
>>>>>>>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>>>>>>>
>>>>>>>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>>>>>>>
>>>>>>>>>>>>> In addition, is the output of this analysis for each
time stamp or for the whole time?
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Shujia
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #52019] use point_stat for multiple forecasts
From: Paul Oldenburg
Time: Thu Dec 15 15:24:40 2011
Shujia,
The commands that I sent you should perform verification on T2 and Q2
(TMP and SPFH). I think you configuration for
these two fields in PointStatConfig_SFC_WRF_sjz is correct.
Paul
On 12/15/2011 03:00 PM, Shujia Zhou via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>
> Paul:
>
> Have you verified my scripts for T2? Is that correct?
>
>
> Thanks,
>
> Shujia
> On Dec 15, 2011, at 12:28 PM, Paul Oldenburg via RT wrote:
>
>> Shujia,
>>
>> I'm sorry for sending this last email prematurely. I meant to add
that soil temperature and moisture observations
>> cannot be extracted or derived from PrepBUFR observations processed
with pb2nc. You will have to find another source of
>> observations if you want to verify those quantities. Please let me
know if you have any other questions.
>>
>> Thanks,
>>
>> Paul
>>
>>
>> On 12/15/2011 10:25 AM, Paul Oldenburg wrote:
>>> Shujia,
>>>
>>> I was able to successfully run point_stat for the model data and
prepbufr data that you sent. First, I had to make the
>>> following change to your PB2NC config file:
>>>
>>> // add SPFH
>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "SPFH"];
>>>
>>>
>>> I used the two commands that included below. Notice that I had to
use the -obs_valid_end flag in the point_stat command
>>> because the PrepBUFR observations are valid at and around
20100602_00 and the model data is valid at 20100602_06.
>>>
>>>
>>>
>>> $MET_BASE/bin/pb2nc \
>>> prepbufr.gdas.20100602.t00z.nr \
>>> prepbufr.gdas.20100602.t00z.nr.nc \
>>> PB2NCConfig_G212_SFC_WRF_sjz \
>>> -v 3
>>>
>>> $MET_BASE/bin/point_stat \
>>> wrf-2010-06-02-06-00-00-PLEV \
>>> prepbufr.gdas.20100602.t00z.nr.nc \
>>> PointStatConfig_SFC_WRF_sjz \
>>> -obs_valid_beg 20100601_22 \
>>> -obs_valid_end 20100602_02 \
>>> -outdir . -v 3
>>>
>>> On 12/15/2011 08:56 AM, Paul Oldenburg wrote:
>>>> Shujia,
>>>>
>>>> I was just informed by my colleague that the file wrf-2010-06-02-
06-00-00-PLEV does appear to the output of p_interp. I
>>>> will analyze this file and respond when I have some results.
Sorry for the confusion.
>>>>
>>>> Thanks,
>>>>
>>>> Paul
>>>>
>>>>
>>>> On 12/15/2011 08:51 AM, Paul Oldenburg wrote:
>>>>> Shujia,
>>>>>
>>>>> The file wrf-2010-06-02-06-00-00-PLEV that you just sent also
appears to be the NetCDF output of WRF, which MET cannot
>>>>> read. MET only supports the following types of model data:
WPP/UPP GRIB output and p_interp NetCDF output. We
>>>>> recommend that users run UPP, since MET has the best support for
GRIB. Do you have access to either WPP/UPP or p_interp?
>>>>>
>>>>> Paul
>>>>>
>>>>>
>>>>> On 12/15/2011 08:48 AM, Shujia Zhou via RT wrote:
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>
>>>>>> Paul:
>>>>>>
>>>>>> I have the data file in pressure level into zhou_data. It is
wrf-2010-06-02-06-00-00-PLEV.
>>>>>> It should work.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> Shujia
>>>>>> On Dec 15, 2011, at 10:24 AM, Paul Oldenburg via RT wrote:
>>>>>>
>>>>>>> Shujia,
>>>>>>>
>>>>>>> The model data file that was included in your tarball,
wrfout_d01_2010-06-02_00:00:00, appears to be the NetCDF output
>>>>>>> of WRF. MET cannot read such files, partly because they are
on eta levels. Do you have access to the WRF Unified Post
>>>>>>> Processor (UPP)? If so, please post-process the WRF NetCDF
files using that tool. Then, MET will be able to read the
>>>>>>> model data. If you have any questions about this, please let
me know.
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> Paul
>>>>>>>
>>>>>>>
>>>>>>> On 12/14/2011 03:17 PM, Shujia Zhou via RT wrote:
>>>>>>>>
>>>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019
>
>>>>>>>>
>>>>>>>> Paul:
>>>>>>>>
>>>>>>>> The tarball is:
>>>>>>>>
>>>>>>>> shujia_met_question.tar.gz
>>>>>>>>
>>>>>>>>
>>>>>>>> Shujia
>>>>>>>> On Dec 14, 2011, at 4:53 PM, Shujia Zhou wrote:
>>>>>>>>
>>>>>>>>> Paul:
>>>>>>>>>
>>>>>>>>> I did put my data along with the scripts into
/incoming/irap/met_help.
>>>>>>>>> (Sorry I forgot putting it into zhou_data.). I would like to
ask
>>>>>>>>> you to see whether my script for T2 is correct or not
although it is
>>>>>>>>> working and the result is reasonable. For the soil_wind dir,
I
>>>>>>>>> put the script for soil and wind, they do not work.
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Shujia
>>>>>>>>> On Dec 14, 2011, at 3:00 PM, Paul Oldenburg via RT wrote:
>>>>>>>>>
>>>>>>>>>> Shujia,
>>>>>>>>>>
>>>>>>>>>> Can you please send me your model forecast data, obs data
and config file? I'll take a look at it and try to build a
>>>>>>>>>> config file for you to use. You can upload data to our FTP
site using these instructions:
>>>>>>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>>
>>>>>>>>>> Paul
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On 12/14/2011 11:49 AM, Shujia Zhou via RT wrote:
>>>>>>>>>>>
>>>>>>>>>>> <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>>>
>>>>>>>>>>> Paul:
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Thank you for your prompt reply.
>>>>>>>>>>>
>>>>>>>>>>> I have two other questions:
>>>>>>>>>>>
>>>>>>>>>>> 1. Do you have a sample PointStatConfig file for surface
wind and soil moisture? Our specification does not work:
>>>>>>>>>>>
>>>>>>>>>>> for wind:
>>>>>>>>>>> fcst_field[] = [ "WIND(0,9,*,*)", "Q2(0,*,*)" ];
>>>>>>>>>>> obs_field[] = [ "WIND/P850", "SPFH/Z2" ];
>>>>>>>>>>>
>>>>>>>>>>> for soil:
>>>>>>>>>>>
>>>>>>>>>>> fcst_field[] = [ "TSLB(0,0,*,*)", "SMOIS(0,0,*,*)",
"Q2(0,*,*)" ];
>>>>>>>>>>>
>>>>>>>>>>> obs_field[] = [ "TSOIL/Z0", "SOILW/Z0", "SPFH/Z2" ];
>>>>>>>>>>>
>>>>>>>>>>> in PB2NCConfig file, we used:
>>>>>>>>>>>
>>>>>>>>>>> message_type[] = ["ANYSFC"];
>>>>>>>>>>>
>>>>>>>>>>> obs_grib_code[] = ["TMP", "WIND", "WDIR", "TSOIL",
"SOILW"];
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> 2. Is we want to have a particular region to compare with
observation, we just need use or create a
>>>>>>>>>>> x.poly file for mask_poly=. Is that right?
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Shujia
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On Dec 14, 2011, at 12:23 PM, Paul Oldenburg via RT wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Shujia,
>>>>>>>>>>>>
>>>>>>>>>>>> The point_stat tool was designed to perform comparisons
for a single valid time. The best approach in your situation
>>>>>>>>>>>> would be to call point_stat once for each forecast valid
time that you list. The obs_valid_beg/end options allow the
>>>>>>>>>>>> user to control the window of observation valid time that
will match the forecast valid time. The fcst_valid/lead
>>>>>>>>>>>> options allow the user to specify which forecast valid
time to verify in an input file that contains multiple valid
>>>>>>>>>>>> times (usually in the form of many lead times for a
single initialization). Does that make sense? Please let me know
>>>>>>>>>>>> if you have any other questions.
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>
>>>>>>>>>>>> Paul
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> On 12/13/2011 04:52 PM, Shujia Zhou via RT wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>> Tue Dec 13 16:52:31 2011: Request 52019 was acted upon.
>>>>>>>>>>>>> Transaction: Ticket created by shujia.zhou-1 at nasa.gov
>>>>>>>>>>>>> Queue: met_help
>>>>>>>>>>>>> Subject: use point_stat for multiple forecasts
>>>>>>>>>>>>> Owner: Nobody
>>>>>>>>>>>>> Requestors: shujia.zhou-1 at nasa.gov
>>>>>>>>>>>>> Status: new
>>>>>>>>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=52019 >
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> We are interested in using point_stat for the multiple
forecasts. For example, 0z, 6z, 12z, 18z in May 29, 2010.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Is this a correct command?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> point_stat forecast_dir observation_dir
PointStatConfig_SFC_WRF –obs_valid_end 2010529_00 –obs_valid_end
2010529_18
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> How about using –fcst_valid and –fcst_lead time?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> In addition, is the output of this analysis for each
time stamp or for the whole time?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Shujia
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>>
>
------------------------------------------------
More information about the Met_help
mailing list