[Met_help] [rt.rap.ucar.edu #63624] History for point_stat evaluation in the upper troposphere
Tressa Fowler via RT
met_help at ucar.edu
Tue Jan 27 08:32:58 MST 2015
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hello,
I am using MET 4.0 patched recently for virt temp in prepbufr - ie. conversion of prepbufr to netcdf is recent.
I am evaluating WRF runs over a month period in summer against prepbufr and am getting unexpected results both for intial time and forecasts.
At lower levels bias in GHT and temperature is within bounds. However, at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is consistently > 0 and often ~ 1K or so. Winds/mixr seem more reasonable.
As recommended I used pinterp with logp interpolation. I tried initial conditions from NAM and GFS and results are very similar. At this point possible problems as I see it can be due to WPS, PINTERP or MET. Do you see these kind of problems in your evaluations?
I could send logs of my MET jobs if that would help. I would appreciate any suggestions on what can be wrong.
Thanks,
Mariusz
Ps: "posterior" is for initial condition (no Data assimilation), "pprior" is 6-hr fcst
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Tressa Fowler
Time: Thu Oct 31 13:07:41 2013
HI Mariusz,
I am sorry your results seem funny. We have not seen similar results,
but we do not use pinterp for our analyses.
Can you use the MET utility plot_data_plane to look at the values in
the temperature and GHT fields at those levels? That may give you some
insight into whether the problem is in WPS or pinterp or MET. If the
values in your files are reasonable, please write back and we will try
to find the cause in MET. If the values are not reasonable, please
write to wrf_help at ucar.edu to see if the issue might be WPS or
pinterp.
Please let us know if we can be of further assistance.
Thanks,
Tressa
On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
> Hello,
>
> I am using MET 4.0 patched recently for virt temp in prepbufr - ie.
> conversion of prepbufr to netcdf is recent.
>
> I am evaluating WRF runs over a month period in summer against
> prepbufr and am getting unexpected results both for intial time
and
> forecasts.
> At lower levels bias in GHT and temperature is within bounds.
However,
> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
> reasonable.
>
> As recommended I used pinterp with logp interpolation. I tried
> initial conditions from NAM and GFS and results are very similar.
> At this point possible problems as I see it can be due to WPS,
> PINTERP or MET. Do you see these kind of problems in your
> evaluations?
>
> I could send logs of my MET jobs if that would help. I would
> appreciate any suggestions on what can be wrong.
>
> Thanks,
> Mariusz
>
> Ps: "posterior" is for initial condition (no Data assimilation),
> "pprior" is 6-hr fcst
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 09:13:39 2013
Tressa,
I cannot find any examples of using plot_data_plane in any of the
manuals incl 4.1. When I try to download 4.1 I cannot find the link
to MET4.1 at
http://www.dtcenter.org/met/users/downloads/index.php
after entering e-mail. There are only links to external libs/utils
etc. (safari, firefox)
Can you help?
Thanks,
Mariusz
On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
> HI Mariusz,
>
> I am sorry your results seem funny. We have not seen similar
results, but we do not use pinterp for our analyses.
>
> Can you use the MET utility plot_data_plane to look at the values in
the temperature and GHT fields at those levels? That may give you some
insight into whether the problem is in WPS or pinterp or MET. If the
values in your files are reasonable, please write back and we will try
to find the cause in MET. If the values are not reasonable, please
write to wrf_help at ucar.edu to see if the issue might be WPS or
pinterp.
>
> Please let us know if we can be of further assistance.
>
> Thanks,
>
> Tressa
>
> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>> Hello,
>>
>> I am using MET 4.0 patched recently for virt temp in prepbufr - ie.
>> conversion of prepbufr to netcdf is recent.
>>
>> I am evaluating WRF runs over a month period in summer against
>> prepbufr and am getting unexpected results both for intial time
and
>> forecasts.
>> At lower levels bias in GHT and temperature is within bounds.
However,
>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>> reasonable.
>>
>> As recommended I used pinterp with logp interpolation. I tried
>> initial conditions from NAM and GFS and results are very similar.
>> At this point possible problems as I see it can be due to WPS,
>> PINTERP or MET. Do you see these kind of problems in your
>> evaluations?
>>
>> I could send logs of my MET jobs if that would help. I would
>> appreciate any suggestions on what can be wrong.
>>
>> Thanks,
>> Mariusz
>>
>> Ps: "posterior" is for initial condition (no Data assimilation),
>> "pprior" is 6-hr fcst
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: John Halley Gotway
Time: Fri Nov 01 09:27:21 2013
Mariusz,
Let me jump in here. So you navigated to this page:
http://www.dtcenter.org/met/users/downloads/index.php
You fill in your email address and then you hit "enter".
You should either be taken to the downloads page including the METv4.1
code or be taken to a registration form if that email address hasn't
been
registered.
Please try again and let me know if it still doesn't work for you.
If not, here's the direct link to the METv4.1 tarball, and be sure to
also
install the latest set of patches:
METv4.1 code:
http://www.dtcenter.org/met/users/downloads/MET_releases/METv4.1.20130522.tar.gz
METv4.1 patches:
http://www.dtcenter.org/met/users/support/known_issues/METv4.1/index.php
Also, some examples of running plot_data_plane can be found in the MET
online tutorial for version 4.1:
http://www.dtcenter.org/met/users/support/online_tutorial/METv4.1/plot_data_plane/index.php
Thanks,
John Halley Gotway
met_help at ucar.edu
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>
> Tressa,
>
> I cannot find any examples of using plot_data_plane in any of the
manuals
> incl 4.1. When I try to download 4.1 I cannot find the link to
MET4.1 at
>
> http://www.dtcenter.org/met/users/downloads/index.php
>
> after entering e-mail. There are only links to external libs/utils
etc.
> (safari, firefox)
>
> Can you help?
>
> Thanks,
> Mariusz
>
>
> On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
>
>> HI Mariusz,
>>
>> I am sorry your results seem funny. We have not seen similar
results,
>> but we do not use pinterp for our analyses.
>>
>> Can you use the MET utility plot_data_plane to look at the values
in the
>> temperature and GHT fields at those levels? That may give you some
>> insight into whether the problem is in WPS or pinterp or MET. If
the
>> values in your files are reasonable, please write back and we will
try
>> to find the cause in MET. If the values are not reasonable, please
write
>> to wrf_help at ucar.edu to see if the issue might be WPS or pinterp.
>>
>> Please let us know if we can be of further assistance.
>>
>> Thanks,
>>
>> Tressa
>>
>> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>>> Hello,
>>>
>>> I am using MET 4.0 patched recently for virt temp in prepbufr -
ie.
>>> conversion of prepbufr to netcdf is recent.
>>>
>>> I am evaluating WRF runs over a month period in summer against
>>> prepbufr and am getting unexpected results both for intial time
and
>>> forecasts.
>>> At lower levels bias in GHT and temperature is within bounds.
However,
>>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>>> reasonable.
>>>
>>> As recommended I used pinterp with logp interpolation. I tried
>>> initial conditions from NAM and GFS and results are very
similar.
>>> At this point possible problems as I see it can be due to WPS,
>>> PINTERP or MET. Do you see these kind of problems in your
>>> evaluations?
>>>
>>> I could send logs of my MET jobs if that would help. I would
>>> appreciate any suggestions on what can be wrong.
>>>
>>> Thanks,
>>> Mariusz
>>>
>>> Ps: "posterior" is for initial condition (no Data assimilation),
>>> "pprior" is 6-hr fcst
>>>
>>
>>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 09:37:49 2013
Thanks John,
that helped. When entering
plot_data_plane $scr1/chem-
var/pagowski/enkf_runs/eval/test_700/met_wrf/wrfout_d01_2012-06-
06_00:00:00_PLEV_posterior TT.ps 'name="TT";level="6";'
plot_data_plane $scr1/chem-
var/pagowski/enkf_runs/eval/test_700/met_wrf/wrfout_d01_2012-06-
06_00:00:00_PLEV_posterior TT.ps 'name="TT";level="6";time="0";'
PinterpFile::data(NcVar *, const LongArray &, DataPlane &, double &)
const -> needed 4 arguments for variable TT, got 3
What am I missing on the command line?
Thanks,
Mariusz
On Nov 1, 2013, at 9:27 AM, John Halley Gotway via RT wrote:
> Mariusz,
>
> Let me jump in here. So you navigated to this page:
> http://www.dtcenter.org/met/users/downloads/index.php
>
> You fill in your email address and then you hit "enter".
>
> You should either be taken to the downloads page including the
METv4.1
> code or be taken to a registration form if that email address hasn't
been
> registered.
>
> Please try again and let me know if it still doesn't work for you.
>
> If not, here's the direct link to the METv4.1 tarball, and be sure
to also
> install the latest set of patches:
> METv4.1 code:
>
http://www.dtcenter.org/met/users/downloads/MET_releases/METv4.1.20130522.tar.gz
> METv4.1 patches:
>
http://www.dtcenter.org/met/users/support/known_issues/METv4.1/index.php
>
> Also, some examples of running plot_data_plane can be found in the
MET
> online tutorial for version 4.1:
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv4.1/plot_data_plane/index.php
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>>
>> Tressa,
>>
>> I cannot find any examples of using plot_data_plane in any of the
manuals
>> incl 4.1. When I try to download 4.1 I cannot find the link to
MET4.1 at
>>
>> http://www.dtcenter.org/met/users/downloads/index.php
>>
>> after entering e-mail. There are only links to external libs/utils
etc.
>> (safari, firefox)
>>
>> Can you help?
>>
>> Thanks,
>> Mariusz
>>
>>
>> On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
>>
>>> HI Mariusz,
>>>
>>> I am sorry your results seem funny. We have not seen similar
results,
>>> but we do not use pinterp for our analyses.
>>>
>>> Can you use the MET utility plot_data_plane to look at the values
in the
>>> temperature and GHT fields at those levels? That may give you some
>>> insight into whether the problem is in WPS or pinterp or MET. If
the
>>> values in your files are reasonable, please write back and we will
try
>>> to find the cause in MET. If the values are not reasonable, please
write
>>> to wrf_help at ucar.edu to see if the issue might be WPS or pinterp.
>>>
>>> Please let us know if we can be of further assistance.
>>>
>>> Thanks,
>>>
>>> Tressa
>>>
>>> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>>>> Hello,
>>>>
>>>> I am using MET 4.0 patched recently for virt temp in prepbufr -
ie.
>>>> conversion of prepbufr to netcdf is recent.
>>>>
>>>> I am evaluating WRF runs over a month period in summer against
>>>> prepbufr and am getting unexpected results both for intial time
and
>>>> forecasts.
>>>> At lower levels bias in GHT and temperature is within bounds.
However,
>>>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>>>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>>>> reasonable.
>>>>
>>>> As recommended I used pinterp with logp interpolation. I tried
>>>> initial conditions from NAM and GFS and results are very
similar.
>>>> At this point possible problems as I see it can be due to WPS,
>>>> PINTERP or MET. Do you see these kind of problems in your
>>>> evaluations?
>>>>
>>>> I could send logs of my MET jobs if that would help. I would
>>>> appreciate any suggestions on what can be wrong.
>>>>
>>>> Thanks,
>>>> Mariusz
>>>>
>>>> Ps: "posterior" is for initial condition (no Data assimilation),
>>>> "pprior" is 6-hr fcst
>>>>
>>>
>>>
>>
>>
>>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 09:46:46 2013
got it, thanks,
MAriusz
On Nov 1, 2013, at 9:27 AM, John Halley Gotway via RT wrote:
> Mariusz,
>
> Let me jump in here. So you navigated to this page:
> http://www.dtcenter.org/met/users/downloads/index.php
>
> You fill in your email address and then you hit "enter".
>
> You should either be taken to the downloads page including the
METv4.1
> code or be taken to a registration form if that email address hasn't
been
> registered.
>
> Please try again and let me know if it still doesn't work for you.
>
> If not, here's the direct link to the METv4.1 tarball, and be sure
to also
> install the latest set of patches:
> METv4.1 code:
>
http://www.dtcenter.org/met/users/downloads/MET_releases/METv4.1.20130522.tar.gz
> METv4.1 patches:
>
http://www.dtcenter.org/met/users/support/known_issues/METv4.1/index.php
>
> Also, some examples of running plot_data_plane can be found in the
MET
> online tutorial for version 4.1:
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv4.1/plot_data_plane/index.php
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>>
>> Tressa,
>>
>> I cannot find any examples of using plot_data_plane in any of the
manuals
>> incl 4.1. When I try to download 4.1 I cannot find the link to
MET4.1 at
>>
>> http://www.dtcenter.org/met/users/downloads/index.php
>>
>> after entering e-mail. There are only links to external libs/utils
etc.
>> (safari, firefox)
>>
>> Can you help?
>>
>> Thanks,
>> Mariusz
>>
>>
>> On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
>>
>>> HI Mariusz,
>>>
>>> I am sorry your results seem funny. We have not seen similar
results,
>>> but we do not use pinterp for our analyses.
>>>
>>> Can you use the MET utility plot_data_plane to look at the values
in the
>>> temperature and GHT fields at those levels? That may give you some
>>> insight into whether the problem is in WPS or pinterp or MET. If
the
>>> values in your files are reasonable, please write back and we will
try
>>> to find the cause in MET. If the values are not reasonable, please
write
>>> to wrf_help at ucar.edu to see if the issue might be WPS or pinterp.
>>>
>>> Please let us know if we can be of further assistance.
>>>
>>> Thanks,
>>>
>>> Tressa
>>>
>>> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>>>> Hello,
>>>>
>>>> I am using MET 4.0 patched recently for virt temp in prepbufr -
ie.
>>>> conversion of prepbufr to netcdf is recent.
>>>>
>>>> I am evaluating WRF runs over a month period in summer against
>>>> prepbufr and am getting unexpected results both for intial time
and
>>>> forecasts.
>>>> At lower levels bias in GHT and temperature is within bounds.
However,
>>>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>>>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>>>> reasonable.
>>>>
>>>> As recommended I used pinterp with logp interpolation. I tried
>>>> initial conditions from NAM and GFS and results are very
similar.
>>>> At this point possible problems as I see it can be due to WPS,
>>>> PINTERP or MET. Do you see these kind of problems in your
>>>> evaluations?
>>>>
>>>> I could send logs of my MET jobs if that would help. I would
>>>> appreciate any suggestions on what can be wrong.
>>>>
>>>> Thanks,
>>>> Mariusz
>>>>
>>>> Ps: "posterior" is for initial condition (no Data assimilation),
>>>> "pprior" is 6-hr fcst
>>>>
>>>
>>>
>>
>>
>>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: John Halley Gotway
Time: Fri Nov 01 10:45:43 2013
Mariusz,
I'm not in the office today, so I can't check to be sure on this...
But I suspect you're just not supplying the "level" information that
plot_data_plane is expecting. Look at the definition of the variable
TT
in your wrfout file:
ncdump -h wrfout_d01_2012-06-06_00:00:00_PLEV_posterior | grep TT
I'm guessing it has 4 dimensions: time, level, x, and y. You need to
tell
plot_data_plane which 2D slice of data to extract and plot. And you
tell
it that by giving it indices for each of the NetCDF dimensions. For
the x
and y dimension, just put in a *. Assuming the dimensions are time,
level, x, and y, in that order, set "level" to "(0,6,*,*)". And run
the
tool like this:
plot_data_plane \
wrfout_d01_2012-06-06_00:00:00_PLEV_posterior \
TT.ps 'name="TT"; level="(0,6,*,*)";'
If I have the dimensions out of order, adjust the level setting
accordingly.
Hope that does it.
Thanks,
John
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>
> Thanks John,
> that helped. When entering
>
> plot_data_plane
> $scr1/chem-
var/pagowski/enkf_runs/eval/test_700/met_wrf/wrfout_d01_2012-06-
06_00:00:00_PLEV_posterior
> TT.ps 'name="TT";level="6";'
>
> plot_data_plane
> $scr1/chem-
var/pagowski/enkf_runs/eval/test_700/met_wrf/wrfout_d01_2012-06-
06_00:00:00_PLEV_posterior
> TT.ps 'name="TT";level="6";time="0";'
>
> PinterpFile::data(NcVar *, const LongArray &, DataPlane &, double &)
const
> -> needed 4 arguments for variable TT, got 3
>
> What am I missing on the command line?
>
> Thanks,
> Mariusz
>
>
> On Nov 1, 2013, at 9:27 AM, John Halley Gotway via RT wrote:
>
>> Mariusz,
>>
>> Let me jump in here. So you navigated to this page:
>> http://www.dtcenter.org/met/users/downloads/index.php
>>
>> You fill in your email address and then you hit "enter".
>>
>> You should either be taken to the downloads page including the
METv4.1
>> code or be taken to a registration form if that email address
hasn't
>> been
>> registered.
>>
>> Please try again and let me know if it still doesn't work for you.
>>
>> If not, here's the direct link to the METv4.1 tarball, and be sure
to
>> also
>> install the latest set of patches:
>> METv4.1 code:
>>
http://www.dtcenter.org/met/users/downloads/MET_releases/METv4.1.20130522.tar.gz
>> METv4.1 patches:
>>
http://www.dtcenter.org/met/users/support/known_issues/METv4.1/index.php
>>
>> Also, some examples of running plot_data_plane can be found in the
MET
>> online tutorial for version 4.1:
>>
http://www.dtcenter.org/met/users/support/online_tutorial/METv4.1/plot_data_plane/index.php
>>
>> Thanks,
>> John Halley Gotway
>> met_help at ucar.edu
>>
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>>>
>>> Tressa,
>>>
>>> I cannot find any examples of using plot_data_plane in any of the
>>> manuals
>>> incl 4.1. When I try to download 4.1 I cannot find the link to
MET4.1
>>> at
>>>
>>> http://www.dtcenter.org/met/users/downloads/index.php
>>>
>>> after entering e-mail. There are only links to external libs/utils
etc.
>>> (safari, firefox)
>>>
>>> Can you help?
>>>
>>> Thanks,
>>> Mariusz
>>>
>>>
>>> On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
>>>
>>>> HI Mariusz,
>>>>
>>>> I am sorry your results seem funny. We have not seen similar
results,
>>>> but we do not use pinterp for our analyses.
>>>>
>>>> Can you use the MET utility plot_data_plane to look at the values
in
>>>> the
>>>> temperature and GHT fields at those levels? That may give you
some
>>>> insight into whether the problem is in WPS or pinterp or MET. If
the
>>>> values in your files are reasonable, please write back and we
will try
>>>> to find the cause in MET. If the values are not reasonable,
please
>>>> write
>>>> to wrf_help at ucar.edu to see if the issue might be WPS or pinterp.
>>>>
>>>> Please let us know if we can be of further assistance.
>>>>
>>>> Thanks,
>>>>
>>>> Tressa
>>>>
>>>> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>>>>> Hello,
>>>>>
>>>>> I am using MET 4.0 patched recently for virt temp in prepbufr -
ie.
>>>>> conversion of prepbufr to netcdf is recent.
>>>>>
>>>>> I am evaluating WRF runs over a month period in summer against
>>>>> prepbufr and am getting unexpected results both for intial time
and
>>>>> forecasts.
>>>>> At lower levels bias in GHT and temperature is within bounds.
>>>>> However,
>>>>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>>>>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>>>>> reasonable.
>>>>>
>>>>> As recommended I used pinterp with logp interpolation. I tried
>>>>> initial conditions from NAM and GFS and results are very
similar.
>>>>> At this point possible problems as I see it can be due to WPS,
>>>>> PINTERP or MET. Do you see these kind of problems in your
>>>>> evaluations?
>>>>>
>>>>> I could send logs of my MET jobs if that would help. I would
>>>>> appreciate any suggestions on what can be wrong.
>>>>>
>>>>> Thanks,
>>>>> Mariusz
>>>>>
>>>>> Ps: "posterior" is for initial condition (no Data assimilation),
>>>>> "pprior" is 6-hr fcst
>>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>>
>>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 11:03:19 2013
John,
thanks, after a while a realized that my level syntax was wrong and
got the figs ok.
Mariusz
On Nov 1, 2013, at 10:45 AM, John Halley Gotway via RT wrote:
> Mariusz,
>
> I'm not in the office today, so I can't check to be sure on this...
>
> But I suspect you're just not supplying the "level" information that
> plot_data_plane is expecting. Look at the definition of the
variable TT
> in your wrfout file:
> ncdump -h wrfout_d01_2012-06-06_00:00:00_PLEV_posterior | grep TT
>
> I'm guessing it has 4 dimensions: time, level, x, and y. You need
to tell
> plot_data_plane which 2D slice of data to extract and plot. And you
tell
> it that by giving it indices for each of the NetCDF dimensions. For
the x
> and y dimension, just put in a *. Assuming the dimensions are time,
> level, x, and y, in that order, set "level" to "(0,6,*,*)". And run
the
> tool like this:
>
> plot_data_plane \
> wrfout_d01_2012-06-06_00:00:00_PLEV_posterior \
> TT.ps 'name="TT"; level="(0,6,*,*)";'
>
> If I have the dimensions out of order, adjust the level setting
accordingly.
>
> Hope that does it.
>
> Thanks,
> John
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>>
>> Thanks John,
>> that helped. When entering
>>
>> plot_data_plane
>> $scr1/chem-
var/pagowski/enkf_runs/eval/test_700/met_wrf/wrfout_d01_2012-06-
06_00:00:00_PLEV_posterior
>> TT.ps 'name="TT";level="6";'
>>
>> plot_data_plane
>> $scr1/chem-
var/pagowski/enkf_runs/eval/test_700/met_wrf/wrfout_d01_2012-06-
06_00:00:00_PLEV_posterior
>> TT.ps 'name="TT";level="6";time="0";'
>>
>> PinterpFile::data(NcVar *, const LongArray &, DataPlane &, double
&) const
>> -> needed 4 arguments for variable TT, got 3
>>
>> What am I missing on the command line?
>>
>> Thanks,
>> Mariusz
>>
>>
>> On Nov 1, 2013, at 9:27 AM, John Halley Gotway via RT wrote:
>>
>>> Mariusz,
>>>
>>> Let me jump in here. So you navigated to this page:
>>> http://www.dtcenter.org/met/users/downloads/index.php
>>>
>>> You fill in your email address and then you hit "enter".
>>>
>>> You should either be taken to the downloads page including the
METv4.1
>>> code or be taken to a registration form if that email address
hasn't
>>> been
>>> registered.
>>>
>>> Please try again and let me know if it still doesn't work for you.
>>>
>>> If not, here's the direct link to the METv4.1 tarball, and be sure
to
>>> also
>>> install the latest set of patches:
>>> METv4.1 code:
>>>
http://www.dtcenter.org/met/users/downloads/MET_releases/METv4.1.20130522.tar.gz
>>> METv4.1 patches:
>>>
http://www.dtcenter.org/met/users/support/known_issues/METv4.1/index.php
>>>
>>> Also, some examples of running plot_data_plane can be found in the
MET
>>> online tutorial for version 4.1:
>>>
http://www.dtcenter.org/met/users/support/online_tutorial/METv4.1/plot_data_plane/index.php
>>>
>>> Thanks,
>>> John Halley Gotway
>>> met_help at ucar.edu
>>>
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>>>>
>>>> Tressa,
>>>>
>>>> I cannot find any examples of using plot_data_plane in any of the
>>>> manuals
>>>> incl 4.1. When I try to download 4.1 I cannot find the link to
MET4.1
>>>> at
>>>>
>>>> http://www.dtcenter.org/met/users/downloads/index.php
>>>>
>>>> after entering e-mail. There are only links to external
libs/utils etc.
>>>> (safari, firefox)
>>>>
>>>> Can you help?
>>>>
>>>> Thanks,
>>>> Mariusz
>>>>
>>>>
>>>> On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
>>>>
>>>>> HI Mariusz,
>>>>>
>>>>> I am sorry your results seem funny. We have not seen similar
results,
>>>>> but we do not use pinterp for our analyses.
>>>>>
>>>>> Can you use the MET utility plot_data_plane to look at the
values in
>>>>> the
>>>>> temperature and GHT fields at those levels? That may give you
some
>>>>> insight into whether the problem is in WPS or pinterp or MET. If
the
>>>>> values in your files are reasonable, please write back and we
will try
>>>>> to find the cause in MET. If the values are not reasonable,
please
>>>>> write
>>>>> to wrf_help at ucar.edu to see if the issue might be WPS or
pinterp.
>>>>>
>>>>> Please let us know if we can be of further assistance.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Tressa
>>>>>
>>>>> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>>>>>> Hello,
>>>>>>
>>>>>> I am using MET 4.0 patched recently for virt temp in prepbufr -
ie.
>>>>>> conversion of prepbufr to netcdf is recent.
>>>>>>
>>>>>> I am evaluating WRF runs over a month period in summer against
>>>>>> prepbufr and am getting unexpected results both for intial time
and
>>>>>> forecasts.
>>>>>> At lower levels bias in GHT and temperature is within bounds.
>>>>>> However,
>>>>>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>>>>>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>>>>>> reasonable.
>>>>>>
>>>>>> As recommended I used pinterp with logp interpolation. I tried
>>>>>> initial conditions from NAM and GFS and results are very
similar.
>>>>>> At this point possible problems as I see it can be due to WPS,
>>>>>> PINTERP or MET. Do you see these kind of problems in your
>>>>>> evaluations?
>>>>>>
>>>>>> I could send logs of my MET jobs if that would help. I would
>>>>>> appreciate any suggestions on what can be wrong.
>>>>>>
>>>>>> Thanks,
>>>>>> Mariusz
>>>>>>
>>>>>> Ps: "posterior" is for initial condition (no Data
assimilation),
>>>>>> "pprior" is 6-hr fcst
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>>
>>
>
>
>
------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 12:01:23 2013
Tressa,
I used plot_data_plane for 250mb and 100mb levels. Nothing in plots
strikes me as apparently wrong.
Please let me know if you think otherwise.
For this area temperature bias at 100mb from point_stat is +1.4K,
geopt bias is +130m
Would that be a possibility if I used your substitute for pinterp?
Thanks,
Mariusz
------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 12:01:23 2013
------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 12:01:23 2013
------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 12:01:23 2013
On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
> HI Mariusz,
>
> I am sorry your results seem funny. We have not seen similar
results, but we do not use pinterp for our analyses.
>
> Can you use the MET utility plot_data_plane to look at the values in
the temperature and GHT fields at those levels? That may give you some
insight into whether the problem is in WPS or pinterp or MET. If the
values in your files are reasonable, please write back and we will try
to find the cause in MET. If the values are not reasonable, please
write to wrf_help at ucar.edu to see if the issue might be WPS or
pinterp.
>
> Please let us know if we can be of further assistance.
>
> Thanks,
>
> Tressa
>
> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>> Hello,
>>
>> I am using MET 4.0 patched recently for virt temp in prepbufr - ie.
>> conversion of prepbufr to netcdf is recent.
>>
>> I am evaluating WRF runs over a month period in summer against
>> prepbufr and am getting unexpected results both for intial time
and
>> forecasts.
>> At lower levels bias in GHT and temperature is within bounds.
However,
>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>> reasonable.
>>
>> As recommended I used pinterp with logp interpolation. I tried
>> initial conditions from NAM and GFS and results are very similar.
>> At this point possible problems as I see it can be due to WPS,
>> PINTERP or MET. Do you see these kind of problems in your
>> evaluations?
>>
>> I could send logs of my MET jobs if that would help. I would
>> appreciate any suggestions on what can be wrong.
>>
>> Thanks,
>> Mariusz
>>
>> Ps: "posterior" is for initial condition (no Data assimilation),
>> "pprior" is 6-hr fcst
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: Tressa Fowler
Time: Fri Nov 01 14:10:32 2013
Hi Mariusz,
I think you should probably send us your files ( model, observations,
and config) so we can figure out what is going on.
Meanwhile, please check in your output statistics for the average
observation and average forecast values at those levels, these should
differ by the amount of your biases, but will tell us whether the
potential problem lies in the obs or forecasts.
To send files, please put them on our anonymous ftp site.
ftp ftp.rap.ucar.edu
username=anonymous
password="your email address"
cd incoming/irap/met_help/pagowski_data
put "your files"
bye
Further, I think you are more in need of debugging support than
statistical support at this point, so I am handing you back to John,
our software engineer.
Thanks,
Tressa
On Nov 1, 2013, at 12:01 PM, Mariusz Pagowski via RT
<met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>
> Tressa,
> I used plot_data_plane for 250mb and 100mb levels. Nothing in plots
strikes me as apparently wrong.
> Please let me know if you think otherwise.
> For this area temperature bias at 100mb from point_stat is +1.4K,
geopt bias is +130m
> Would that be a possibility if I used your substitute for pinterp?
> Thanks,
> Mariusz
>
>
> <TT_6.pdf>
>
> <GHT_6.pdf>
>
> <TT_7.pdf>
> <GHT_7.pdf>
>
> On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
>
>> HI Mariusz,
>>
>> I am sorry your results seem funny. We have not seen similar
results, but we do not use pinterp for our analyses.
>>
>> Can you use the MET utility plot_data_plane to look at the values
in the temperature and GHT fields at those levels? That may give you
some insight into whether the problem is in WPS or pinterp or MET. If
the values in your files are reasonable, please write back and we will
try to find the cause in MET. If the values are not reasonable, please
write to wrf_help at ucar.edu to see if the issue might be WPS or
pinterp.
>>
>> Please let us know if we can be of further assistance.
>>
>> Thanks,
>>
>> Tressa
>>
>> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>>> Hello,
>>>
>>> I am using MET 4.0 patched recently for virt temp in prepbufr -
ie.
>>> conversion of prepbufr to netcdf is recent.
>>>
>>> I am evaluating WRF runs over a month period in summer against
>>> prepbufr and am getting unexpected results both for intial time
and
>>> forecasts.
>>> At lower levels bias in GHT and temperature is within bounds.
However,
>>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>>> reasonable.
>>>
>>> As recommended I used pinterp with logp interpolation. I tried
>>> initial conditions from NAM and GFS and results are very similar.
>>> At this point possible problems as I see it can be due to WPS,
>>> PINTERP or MET. Do you see these kind of problems in your
>>> evaluations?
>>>
>>> I could send logs of my MET jobs if that would help. I would
>>> appreciate any suggestions on what can be wrong.
>>>
>>> Thanks,
>>> Mariusz
>>>
>>> Ps: "posterior" is for initial condition (no Data assimilation),
>>> "pprior" is 6-hr fcst
>>>
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #63624] point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Fri Nov 01 15:08:25 2013
Tressa,
I very much appreciate your help. I posted the files on ftp.
Line from my point_stat output reads like below.
Mariusz
FCST_VALID_BEG,FCST_VAR,FCST_LEV,OBS_VAR,OBS_LEV,OBTYPE,TOTAL,
FBAR,FSTDEV,OBAR, OSTDEV,PR_CORR, ME, BCMSE,RMSE
20120606_000000,TT,P100,TMP,P100,ADPUPA,41,
211.36484,5.82217,209.91585, 6.29681,0.99120, 1.44898,
0.84922,1.71720
On Nov 1, 2013, at 2:10 PM, Tressa Fowler via RT wrote:
> Hi Mariusz,
>
> I think you should probably send us your files ( model,
observations, and config) so we can figure out what is going on.
>
> Meanwhile, please check in your output statistics for the average
observation and average forecast values at those levels, these should
differ by the amount of your biases, but will tell us whether the
potential problem lies in the obs or forecasts.
>
> To send files, please put them on our anonymous ftp site.
>
> ftp ftp.rap.ucar.edu
> username=anonymous
> password="your email address"
> cd incoming/irap/met_help/pagowski_data
> put "your files"
> bye
>
>
> Further, I think you are more in need of debugging support than
statistical support at this point, so I am handing you back to John,
our software engineer.
>
> Thanks,
>
> Tressa
>
> On Nov 1, 2013, at 12:01 PM, Mariusz Pagowski via RT
<met_help at ucar.edu> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=63624 >
>>
>> Tressa,
>> I used plot_data_plane for 250mb and 100mb levels. Nothing in plots
strikes me as apparently wrong.
>> Please let me know if you think otherwise.
>> For this area temperature bias at 100mb from point_stat is +1.4K,
geopt bias is +130m
>> Would that be a possibility if I used your substitute for pinterp?
>> Thanks,
>> Mariusz
>>
>>
>> <TT_6.pdf>
>>
>> <GHT_6.pdf>
>>
>> <TT_7.pdf>
>> <GHT_7.pdf>
>>
>> On Oct 31, 2013, at 1:07 PM, Tressa Fowler via RT wrote:
>>
>>> HI Mariusz,
>>>
>>> I am sorry your results seem funny. We have not seen similar
results, but we do not use pinterp for our analyses.
>>>
>>> Can you use the MET utility plot_data_plane to look at the values
in the temperature and GHT fields at those levels? That may give you
some insight into whether the problem is in WPS or pinterp or MET. If
the values in your files are reasonable, please write back and we will
try to find the cause in MET. If the values are not reasonable, please
write to wrf_help at ucar.edu to see if the issue might be WPS or
pinterp.
>>>
>>> Please let us know if we can be of further assistance.
>>>
>>> Thanks,
>>>
>>> Tressa
>>>
>>> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>>>> Hello,
>>>>
>>>> I am using MET 4.0 patched recently for virt temp in prepbufr -
ie.
>>>> conversion of prepbufr to netcdf is recent.
>>>>
>>>> I am evaluating WRF runs over a month period in summer against
>>>> prepbufr and am getting unexpected results both for intial time
and
>>>> forecasts.
>>>> At lower levels bias in GHT and temperature is within bounds.
However,
>>>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>>>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>>>> reasonable.
>>>>
>>>> As recommended I used pinterp with logp interpolation. I tried
>>>> initial conditions from NAM and GFS and results are very similar.
>>>> At this point possible problems as I see it can be due to WPS,
>>>> PINTERP or MET. Do you see these kind of problems in your
>>>> evaluations?
>>>>
>>>> I could send logs of my MET jobs if that would help. I would
>>>> appreciate any suggestions on what can be wrong.
>>>>
>>>> Thanks,
>>>> Mariusz
>>>>
>>>> Ps: "posterior" is for initial condition (no Data assimilation),
>>>> "pprior" is 6-hr fcst
>>>>
>>>
>>>
>>
>
>
>
------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Tressa Fowler
Time: Wed Nov 06 14:17:51 2013
Hi Mariusz,
I am sorry, when I saw you reporting temperature bias of 1.4K, I
thought you meant 1400 degrees, not 1.4 Kelvin. We looked at your
data, looks normal to us. I am attaching a plot showing some results
we have for temperature bias which shows biases similar to yours. We
do not examine geopotential height. Thus, we feel pretty sure there is
no software or interpolation issue. Please let us know if we can help
you further.
Thanks,
Tressa
On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
> Hello,
>
> I am using MET 4.0 patched recently for virt temp in prepbufr - ie.
> conversion of prepbufr to netcdf is recent.
>
> I am evaluating WRF runs over a month period in summer against
> prepbufr and am getting unexpected results both for intial time
and
> forecasts.
> At lower levels bias in GHT and temperature is within bounds.
However,
> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
> reasonable.
>
> As recommended I used pinterp with logp interpolation. I tried
> initial conditions from NAM and GFS and results are very similar.
> At this point possible problems as I see it can be due to WPS,
> PINTERP or MET. Do you see these kind of problems in your
> evaluations?
>
> I could send logs of my MET jobs if that would help. I would
> appreciate any suggestions on what can be wrong.
>
> Thanks,
> Mariusz
>
> Ps: "posterior" is for initial condition (no Data assimilation),
> "pprior" is 6-hr fcst
>
------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Thu Nov 07 12:58:48 2013
Tressa,
Biases in geopt and temperature are consistent as delta_h ~ T but the
reason is by no means obvious to me. The plots that I sent are one-
month averages and I still think that daily biases of these magnitudes
are unusual and nearly always positive. I find that hard to
comprehend.
Eg. please compare them to ARW-wrf run at GSD for a similar period
last summer. I don't know details of verification but they use WPS to
initialize - these are just initial conditions, not forecasts.
Would both NAM and GFS have such positive biases?
I might be able to find other r-t verifications as another example but
people usually don't post them online.
Thanks,
Mariusz
------------------------------------------------
Subject: point_stat evaluation in the upper troposphere
From: Mariusz Pagowski
Time: Thu Nov 07 12:58:48 2013
On Nov 6, 2013, at 2:17 PM, Tressa Fowler via RT wrote:
> Hi Mariusz,
>
> I am sorry, when I saw you reporting temperature bias of 1.4K, I
thought you meant 1400 degrees, not 1.4 Kelvin. We looked at your
data, looks normal to us. I am attaching a plot showing some results
we have for temperature bias which shows biases similar to yours. We
do not examine geopotential height. Thus, we feel pretty sure there is
no software or interpolation issue. Please let us know if we can help
you further.
>
> Thanks,
>
> Tressa
>
> On Mon Oct 28 12:32:42 2013, Mariusz.Pagowski at noaa.gov wrote:
>> Hello,
>>
>> I am using MET 4.0 patched recently for virt temp in prepbufr - ie.
>> conversion of prepbufr to netcdf is recent.
>>
>> I am evaluating WRF runs over a month period in summer against
>> prepbufr and am getting unexpected results both for intial time
and
>> forecasts.
>> At lower levels bias in GHT and temperature is within bounds.
However,
>> at 250 and 100 mbs the bias (ME) is unexpectedly high ie. is
>> consistently > 0 and often ~ 1K or so. Winds/mixr seem more
>> reasonable.
>>
>> As recommended I used pinterp with logp interpolation. I tried
>> initial conditions from NAM and GFS and results are very similar.
>> At this point possible problems as I see it can be due to WPS,
>> PINTERP or MET. Do you see these kind of problems in your
>> evaluations?
>>
>> I could send logs of my MET jobs if that would help. I would
>> appreciate any suggestions on what can be wrong.
>>
>> Thanks,
>> Mariusz
>>
>> Ps: "posterior" is for initial condition (no Data assimilation),
>> "pprior" is 6-hr fcst
>>
>
>
>
> <TMP_ME_CONUS_f00_BILIN.png>
------------------------------------------------
More information about the Met_help
mailing list