[Met_help] [rt.rap.ucar.edu #54901] History for Use of MET directly with wrfout history file
John Halley Gotway via RT
met_help at ucar.edu
Fri Mar 9 13:42:17 MST 2012
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi,
If I'm not going to use the vector fields but only one particular scalar
field (precipitation), do I still need to post-process/destagger the wrfout
file using WPP/UPP? Can I use it directly in MET?
Thanks,
Raizan
>
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #54901] Use of MET directly with wrfout history file
From: John Halley Gotway
Time: Mon Mar 05 11:37:03 2012
Raizan,
Unfortunately, yes. MET is unable to read the NetCDF output of WRF
directly. Instead, you must first post-process it using either the
Unified PostProcessor (UPP) or the pinterp tool (for ARW only).
We strongly recommend using UPP since MET's support is most robust
for the GRIB1 format that UPP creates.
Thanks,
John Halley Gotway
met_help at ucar.edu
On 03/05/2012 11:23 AM, Zan Rahmat via RT wrote:
>
> Mon Mar 05 11:23:24 2012: Request 54901 was acted upon.
> Transaction: Ticket created by raizan.rahmat at gmail.com
> Queue: met_help
> Subject: Use of MET directly with wrfout history file
> Owner: Nobody
> Requestors: raizan.rahmat at gmail.com
> Status: new
> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
>
>
> Hi,
>
> If I'm not going to use the vector fields but only one particular
scalar
> field (precipitation), do I still need to post-process/destagger the
wrfout
> file using WPP/UPP? Can I use it directly in MET?
>
> Thanks,
> Raizan
>
>
>>
------------------------------------------------
Subject: Use of MET directly with wrfout history file
From: Zan Rahmat
Time: Mon Mar 05 12:08:17 2012
Thanks John,
1. I will give pinterp tool a try too (having problems with UPP tool
due to
memory fault, a colleague of yours have requested me to ask wrfhelp on
that
instead).
2. If I have a precipitation analyses field (observation) in binary
(not in
GRIB), and then I convert it to netcdf, can I use it in MET
straightaway?
Or are there any particular netcdf format that MET requires for the
gridded
observation data?
Thanks,
Raizan
On 5 March 2012 18:37, John Halley Gotway via RT <met_help at ucar.edu>
wrote:
> Raizan,
>
> Unfortunately, yes. MET is unable to read the NetCDF output of WRF
> directly. Instead, you must first post-process it using either the
Unified
> PostProcessor (UPP) or the pinterp tool (for ARW only).
> We strongly recommend using UPP since MET's support is most robust
for
> the GRIB1 format that UPP creates.
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
> On 03/05/2012 11:23 AM, Zan Rahmat via RT wrote:
> >
> > Mon Mar 05 11:23:24 2012: Request 54901 was acted upon.
> > Transaction: Ticket created by raizan.rahmat at gmail.com
> > Queue: met_help
> > Subject: Use of MET directly with wrfout history file
> > Owner: Nobody
> > Requestors: raizan.rahmat at gmail.com
> > Status: new
> > Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
> >
> >
> > Hi,
> >
> > If I'm not going to use the vector fields but only one particular
scalar
> > field (precipitation), do I still need to post-process/destagger
the
> wrfout
> > file using WPP/UPP? Can I use it directly in MET?
> >
> > Thanks,
> > Raizan
> >
> >
> >>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #54901] Use of MET directly with wrfout history file
From: John Halley Gotway
Time: Mon Mar 05 12:21:30 2012
Raizan,
Yes, it's often getting data in the right format that's the most
difficult part of verification!
Can you tell me the source of the binary precipitation analysis field
you're using? If it's a common format, it's possible that I could
recommend a tool for reformatting it into something MET could
read. If not, then yes, you could put it into a NetCDF format.
You'll basically want to make it look like the output of the
pcp_combine tool. That's an example of the NetCDF format that MET is
currently set up to read.
However, be aware that your gridded forecast and observation data will
need to be put on a common grid before verifying. If they're both in
NetCDF format, and they are not already on the same grid,
regridding is difficult. But if you're able to get UPP running and
have forecast data in GRIB format, the copygb utility makes it easy to
regrid GRIB files.
Thanks,
John
On 03/05/2012 12:08 PM, Zan Rahmat via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
>
> Thanks John,
>
> 1. I will give pinterp tool a try too (having problems with UPP tool
due to
> memory fault, a colleague of yours have requested me to ask wrfhelp
on that
> instead).
>
> 2. If I have a precipitation analyses field (observation) in binary
(not in
> GRIB), and then I convert it to netcdf, can I use it in MET
straightaway?
> Or are there any particular netcdf format that MET requires for the
gridded
> observation data?
>
> Thanks,
> Raizan
>
> On 5 March 2012 18:37, John Halley Gotway via RT<met_help at ucar.edu>
wrote:
>
>> Raizan,
>>
>> Unfortunately, yes. MET is unable to read the NetCDF output of WRF
>> directly. Instead, you must first post-process it using either the
Unified
>> PostProcessor (UPP) or the pinterp tool (for ARW only).
>> We strongly recommend using UPP since MET's support is most
robust for
>> the GRIB1 format that UPP creates.
>>
>> Thanks,
>> John Halley Gotway
>> met_help at ucar.edu
>>
>> On 03/05/2012 11:23 AM, Zan Rahmat via RT wrote:
>>>
>>> Mon Mar 05 11:23:24 2012: Request 54901 was acted upon.
>>> Transaction: Ticket created by raizan.rahmat at gmail.com
>>> Queue: met_help
>>> Subject: Use of MET directly with wrfout history file
>>> Owner: Nobody
>>> Requestors: raizan.rahmat at gmail.com
>>> Status: new
>>> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
>>>
>>>
>>> Hi,
>>>
>>> If I'm not going to use the vector fields but only one particular
scalar
>>> field (precipitation), do I still need to post-process/destagger
the
>> wrfout
>>> file using WPP/UPP? Can I use it directly in MET?
>>>
>>> Thanks,
>>> Raizan
>>>
>>>
>>>>
>>
>>
------------------------------------------------
Subject: Use of MET directly with wrfout history file
From: Zan Rahmat
Time: Mon Mar 05 12:52:24 2012
Dear John,
Thanks.That's very helpful!
If I may just ask 2 more questions please, just to be very sure and
specific.
If I manage to get UPP running and attempt to regrid the wrfout
history
file using copygb eventually...
1. Is using copygb necessary for the purpose of comparing obs and
model
output produced by ARW?
I'm using ARW, not NMM. I have read that for wrfout produced by ARW,
it is
not necessary to run the copygb - my impression is this is only true
for
plotting purposes only. For verification, regardless if the wrfout is
produced by ARW or NMM, copygb is still a necessary step to regrid the
ARW
model output file. Is my conclusion correct please?
2. Are you saying that as long as I manage to get the binary
precipitation
data into a netCDF format that looks like the output of pcp_combine,
there
is no further need to regrid it to match the regridded ARW wrfout
history
file? Suffice to only copygb the model output above to regrid the
model
output?
3. I'm using NIMROD product data which are in a non-standard binary
format
described in the (
http://badc.nerc.ac.uk/view/badc.nerc.ac.uk__ATOM__dataent_nimrod).
Any
tools suitable for this?
Thanks,
On 5 March 2012 19:21, John Halley Gotway via RT <met_help at ucar.edu>
wrote:
> Raizan,
>
> Yes, it's often getting data in the right format that's the most
difficult
> part of verification!
>
> Can you tell me the source of the binary precipitation analysis
field
> you're using? If it's a common format, it's possible that I could
> recommend a tool for reformatting it into something MET could
> read. If not, then yes, you could put it into a NetCDF format.
You'll
> basically want to make it look like the output of the pcp_combine
tool.
> That's an example of the NetCDF format that MET is
> currently set up to read.
>
> However, be aware that your gridded forecast and observation data
will
> need to be put on a common grid before verifying. If they're both
in
> NetCDF format, and they are not already on the same grid,
> regridding is difficult. But if you're able to get UPP running and
have
> forecast data in GRIB format, the copygb utility makes it easy to
regrid
> GRIB files.
>
> Thanks,
> John
>
> On 03/05/2012 12:08 PM, Zan Rahmat via RT wrote:
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
> >
> > Thanks John,
> >
> > 1. I will give pinterp tool a try too (having problems with UPP
tool due
> to
> > memory fault, a colleague of yours have requested me to ask
wrfhelp on
> that
> > instead).
> >
> > 2. If I have a precipitation analyses field (observation) in
binary (not
> in
> > GRIB), and then I convert it to netcdf, can I use it in MET
straightaway?
> > Or are there any particular netcdf format that MET requires for
the
> gridded
> > observation data?
> >
> > Thanks,
> > Raizan
> >
> > On 5 March 2012 18:37, John Halley Gotway via
RT<met_help at ucar.edu>
> wrote:
> >
> >> Raizan,
> >>
> >> Unfortunately, yes. MET is unable to read the NetCDF output of
WRF
> >> directly. Instead, you must first post-process it using either
the
> Unified
> >> PostProcessor (UPP) or the pinterp tool (for ARW only).
> >> We strongly recommend using UPP since MET's support is most
robust for
> >> the GRIB1 format that UPP creates.
> >>
> >> Thanks,
> >> John Halley Gotway
> >> met_help at ucar.edu
> >>
> >> On 03/05/2012 11:23 AM, Zan Rahmat via RT wrote:
> >>>
> >>> Mon Mar 05 11:23:24 2012: Request 54901 was acted upon.
> >>> Transaction: Ticket created by raizan.rahmat at gmail.com
> >>> Queue: met_help
> >>> Subject: Use of MET directly with wrfout history file
> >>> Owner: Nobody
> >>> Requestors: raizan.rahmat at gmail.com
> >>> Status: new
> >>> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901
> >
> >>>
> >>>
> >>> Hi,
> >>>
> >>> If I'm not going to use the vector fields but only one
particular
> scalar
> >>> field (precipitation), do I still need to post-process/destagger
the
> >> wrfout
> >>> file using WPP/UPP? Can I use it directly in MET?
> >>>
> >>> Thanks,
> >>> Raizan
> >>>
> >>>
> >>>>
> >>
> >>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #54901] Use of MET directly with wrfout history file
From: John Halley Gotway
Time: Mon Mar 05 13:25:48 2012
Raizan,
Hopefully the following information will clarify...
- MET can read the GRIB output of UPP or the NetCDF output of the
pinterp tool.
- All of the MET grid-to-grid comparison tools require that the
forecast and observation datasets be on a common grid. And it is the
user's responsibility to put them on a common grid.
So this is not an issue of NMM vs ARW. My warning to you is that if
you end up with 2 NetCDF files (using pinterp), putting them on a
common grid may be difficult. However, if you have one GRIB file
(from UPP) and one NetCDF file (NIMROD data reformatted into NetCDF),
you could use the copygb utility to regrid your GRIB model data onto
the NIMROD domain.
copygb is a nice utility that enables you to regrid GRIB files. It is
up to you to decide how you'd like to proceed, but here are the steps
I'd recommend:
(1) Post-process your WRF output using UPP to generate GRIB1 files.
(2) Write a script to convert the NIMROD data from binary into a
NetCDF format that MET can read. Make it look like the output of the
PCP-Combine tool.
(3) Run your GRIB1 model files through copygb to regrid them to the
NIMROD domain.
(4) Run the MET grid-to-grid tools (grid_stat, MODE, and wavelet-stat)
to compare your model output to the NIMROD data.
Unfortunately, I don't have any code available for handling the NIMROD
binary data format. Sorry I can't be of more help.
But if you run into any problems getting the NIMROD data into a NetCDF
format that MET can handle, just let us know.
Thanks,
John
On 03/05/2012 12:52 PM, Zan Rahmat via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
>
> Dear John,
>
> Thanks.That's very helpful!
>
> If I may just ask 2 more questions please, just to be very sure and
> specific.
>
> If I manage to get UPP running and attempt to regrid the wrfout
history
> file using copygb eventually...
>
> 1. Is using copygb necessary for the purpose of comparing obs and
model
> output produced by ARW?
> I'm using ARW, not NMM. I have read that for wrfout produced by ARW,
it is
> not necessary to run the copygb - my impression is this is only true
for
> plotting purposes only. For verification, regardless if the wrfout
is
> produced by ARW or NMM, copygb is still a necessary step to regrid
the ARW
> model output file. Is my conclusion correct please?
>
> 2. Are you saying that as long as I manage to get the binary
precipitation
> data into a netCDF format that looks like the output of pcp_combine,
there
> is no further need to regrid it to match the regridded ARW wrfout
history
> file? Suffice to only copygb the model output above to regrid the
model
> output?
>
> 3. I'm using NIMROD product data which are in a non-standard binary
format
> described in the (
> http://badc.nerc.ac.uk/view/badc.nerc.ac.uk__ATOM__dataent_nimrod).
Any
> tools suitable for this?
>
> Thanks,
>
>
> On 5 March 2012 19:21, John Halley Gotway via RT<met_help at ucar.edu>
wrote:
>
>> Raizan,
>>
>> Yes, it's often getting data in the right format that's the most
difficult
>> part of verification!
>>
>> Can you tell me the source of the binary precipitation analysis
field
>> you're using? If it's a common format, it's possible that I could
>> recommend a tool for reformatting it into something MET could
>> read. If not, then yes, you could put it into a NetCDF format.
You'll
>> basically want to make it look like the output of the pcp_combine
tool.
>> That's an example of the NetCDF format that MET is
>> currently set up to read.
>>
>> However, be aware that your gridded forecast and observation data
will
>> need to be put on a common grid before verifying. If they're both
in
>> NetCDF format, and they are not already on the same grid,
>> regridding is difficult. But if you're able to get UPP running and
have
>> forecast data in GRIB format, the copygb utility makes it easy to
regrid
>> GRIB files.
>>
>> Thanks,
>> John
>>
>> On 03/05/2012 12:08 PM, Zan Rahmat via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
>>>
>>> Thanks John,
>>>
>>> 1. I will give pinterp tool a try too (having problems with UPP
tool due
>> to
>>> memory fault, a colleague of yours have requested me to ask
wrfhelp on
>> that
>>> instead).
>>>
>>> 2. If I have a precipitation analyses field (observation) in
binary (not
>> in
>>> GRIB), and then I convert it to netcdf, can I use it in MET
straightaway?
>>> Or are there any particular netcdf format that MET requires for
the
>> gridded
>>> observation data?
>>>
>>> Thanks,
>>> Raizan
>>>
>>> On 5 March 2012 18:37, John Halley Gotway via
RT<met_help at ucar.edu>
>> wrote:
>>>
>>>> Raizan,
>>>>
>>>> Unfortunately, yes. MET is unable to read the NetCDF output of
WRF
>>>> directly. Instead, you must first post-process it using either
the
>> Unified
>>>> PostProcessor (UPP) or the pinterp tool (for ARW only).
>>>> We strongly recommend using UPP since MET's support is most
robust for
>>>> the GRIB1 format that UPP creates.
>>>>
>>>> Thanks,
>>>> John Halley Gotway
>>>> met_help at ucar.edu
>>>>
>>>> On 03/05/2012 11:23 AM, Zan Rahmat via RT wrote:
>>>>>
>>>>> Mon Mar 05 11:23:24 2012: Request 54901 was acted upon.
>>>>> Transaction: Ticket created by raizan.rahmat at gmail.com
>>>>> Queue: met_help
>>>>> Subject: Use of MET directly with wrfout history file
>>>>> Owner: Nobody
>>>>> Requestors: raizan.rahmat at gmail.com
>>>>> Status: new
>>>>> Ticket<URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901
>>>
>>>>>
>>>>>
>>>>> Hi,
>>>>>
>>>>> If I'm not going to use the vector fields but only one
particular
>> scalar
>>>>> field (precipitation), do I still need to post-process/destagger
the
>>>> wrfout
>>>>> file using WPP/UPP? Can I use it directly in MET?
>>>>>
>>>>> Thanks,
>>>>> Raizan
>>>>>
>>>>>
>>>>>>
>>>>
>>>>
>>
>>
------------------------------------------------
Subject: Use of MET directly with wrfout history file
From: Zan Rahmat
Time: Mon Mar 05 13:32:08 2012
Thanks John!
It is very clear now. I will give the steps you outlined a try. Hope I
can
get UPP working properly first.
Warm Regards,
Raizan
On 5 March 2012 20:25, John Halley Gotway via RT <met_help at ucar.edu>
wrote:
> Raizan,
>
> Hopefully the following information will clarify...
>
> - MET can read the GRIB output of UPP or the NetCDF output of the
pinterp
> tool.
> - All of the MET grid-to-grid comparison tools require that the
forecast
> and observation datasets be on a common grid. And it is the user's
> responsibility to put them on a common grid.
>
> So this is not an issue of NMM vs ARW. My warning to you is that if
you
> end up with 2 NetCDF files (using pinterp), putting them on a common
grid
> may be difficult. However, if you have one GRIB file
> (from UPP) and one NetCDF file (NIMROD data reformatted into
NetCDF), you
> could use the copygb utility to regrid your GRIB model data onto the
NIMROD
> domain.
>
> copygb is a nice utility that enables you to regrid GRIB files. It
is up
> to you to decide how you'd like to proceed, but here are the steps
I'd
> recommend:
>
> (1) Post-process your WRF output using UPP to generate GRIB1 files.
> (2) Write a script to convert the NIMROD data from binary into a
NetCDF
> format that MET can read. Make it look like the output of the PCP-
Combine
> tool.
> (3) Run your GRIB1 model files through copygb to regrid them to the
NIMROD
> domain.
> (4) Run the MET grid-to-grid tools (grid_stat, MODE, and wavelet-
stat) to
> compare your model output to the NIMROD data.
>
> Unfortunately, I don't have any code available for handling the
NIMROD
> binary data format. Sorry I can't be of more help.
>
> But if you run into any problems getting the NIMROD data into a
NetCDF
> format that MET can handle, just let us know.
>
> Thanks,
> John
>
>
>
> On 03/05/2012 12:52 PM, Zan Rahmat via RT wrote:
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
> >
> > Dear John,
> >
> > Thanks.That's very helpful!
> >
> > If I may just ask 2 more questions please, just to be very sure
and
> > specific.
> >
> > If I manage to get UPP running and attempt to regrid the wrfout
history
> > file using copygb eventually...
> >
> > 1. Is using copygb necessary for the purpose of comparing obs and
model
> > output produced by ARW?
> > I'm using ARW, not NMM. I have read that for wrfout produced by
ARW, it
> is
> > not necessary to run the copygb - my impression is this is only
true for
> > plotting purposes only. For verification, regardless if the wrfout
is
> > produced by ARW or NMM, copygb is still a necessary step to regrid
the
> ARW
> > model output file. Is my conclusion correct please?
> >
> > 2. Are you saying that as long as I manage to get the binary
> precipitation
> > data into a netCDF format that looks like the output of
pcp_combine,
> there
> > is no further need to regrid it to match the regridded ARW wrfout
history
> > file? Suffice to only copygb the model output above to regrid the
model
> > output?
> >
> > 3. I'm using NIMROD product data which are in a non-standard
binary
> format
> > described in the (
> >
http://badc.nerc.ac.uk/view/badc.nerc.ac.uk__ATOM__dataent_nimrod).
Any
> > tools suitable for this?
> >
> > Thanks,
> >
> >
> > On 5 March 2012 19:21, John Halley Gotway via
RT<met_help at ucar.edu>
> wrote:
> >
> >> Raizan,
> >>
> >> Yes, it's often getting data in the right format that's the most
> difficult
> >> part of verification!
> >>
> >> Can you tell me the source of the binary precipitation analysis
field
> >> you're using? If it's a common format, it's possible that I
could
> >> recommend a tool for reformatting it into something MET could
> >> read. If not, then yes, you could put it into a NetCDF format.
You'll
> >> basically want to make it look like the output of the pcp_combine
tool.
> >> That's an example of the NetCDF format that MET is
> >> currently set up to read.
> >>
> >> However, be aware that your gridded forecast and observation data
will
> >> need to be put on a common grid before verifying. If they're
both in
> >> NetCDF format, and they are not already on the same grid,
> >> regridding is difficult. But if you're able to get UPP running
and have
> >> forecast data in GRIB format, the copygb utility makes it easy to
regrid
> >> GRIB files.
> >>
> >> Thanks,
> >> John
> >>
> >> On 03/05/2012 12:08 PM, Zan Rahmat via RT wrote:
> >>>
> >>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901>
> >>>
> >>> Thanks John,
> >>>
> >>> 1. I will give pinterp tool a try too (having problems with UPP
tool
> due
> >> to
> >>> memory fault, a colleague of yours have requested me to ask
wrfhelp on
> >> that
> >>> instead).
> >>>
> >>> 2. If I have a precipitation analyses field (observation) in
binary
> (not
> >> in
> >>> GRIB), and then I convert it to netcdf, can I use it in MET
> straightaway?
> >>> Or are there any particular netcdf format that MET requires for
the
> >> gridded
> >>> observation data?
> >>>
> >>> Thanks,
> >>> Raizan
> >>>
> >>> On 5 March 2012 18:37, John Halley Gotway via
RT<met_help at ucar.edu>
> >> wrote:
> >>>
> >>>> Raizan,
> >>>>
> >>>> Unfortunately, yes. MET is unable to read the NetCDF output of
WRF
> >>>> directly. Instead, you must first post-process it using either
the
> >> Unified
> >>>> PostProcessor (UPP) or the pinterp tool (for ARW only).
> >>>> We strongly recommend using UPP since MET's support is most
robust
> for
> >>>> the GRIB1 format that UPP creates.
> >>>>
> >>>> Thanks,
> >>>> John Halley Gotway
> >>>> met_help at ucar.edu
> >>>>
> >>>> On 03/05/2012 11:23 AM, Zan Rahmat via RT wrote:
> >>>>>
> >>>>> Mon Mar 05 11:23:24 2012: Request 54901 was acted upon.
> >>>>> Transaction: Ticket created by raizan.rahmat at gmail.com
> >>>>> Queue: met_help
> >>>>> Subject: Use of MET directly with wrfout history file
> >>>>> Owner: Nobody
> >>>>> Requestors: raizan.rahmat at gmail.com
> >>>>> Status: new
> >>>>> Ticket<URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=54901
> >>>
> >>>>>
> >>>>>
> >>>>> Hi,
> >>>>>
> >>>>> If I'm not going to use the vector fields but only one
particular
> >> scalar
> >>>>> field (precipitation), do I still need to post-
process/destagger the
> >>>> wrfout
> >>>>> file using WPP/UPP? Can I use it directly in MET?
> >>>>>
> >>>>> Thanks,
> >>>>> Raizan
> >>>>>
> >>>>>
> >>>>>>
> >>>>
> >>>>
> >>
> >>
>
>
------------------------------------------------
More information about the Met_help
mailing list