[Met_help] [rt.rap.ucar.edu #71614] History for ascii2nc followup
John Halley Gotway via RT
met_help at ucar.edu
Mon May 4 09:44:41 MDT 2015
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hello,
I have observations of dew point and 6-hr accumulated precipitation, write them out to ascii file setting codes 17 and 61, respectively, and convert to prepbufer with ascii2nc.
I later want to use pointstat* for WRF evaluation against these data. There are no corresponding variables in WRF output. Do I need to explicitly calculate
mixing ratio from Td and some other form of precip from these observations so they correspond to WRF output variables? Or there is some trick that is available in pointstat configuration that would do it for me?
Thanks,
Mariusz
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: ascii2nc followup
From: John Halley Gotway
Time: Fri Apr 24 10:12:24 2015
Mariusz,
While the PB2NC tool has the ability to derive observation types from
input
point observations (i.e. convert specific humidity and temperature to
relative humidity), the ASCII2NC and MADIS2NC tools do not. This has
come
up in the past, and we would like to add this functionality but
haven't
found the time or funding to do so yet.
Your choices are...
- Write a script to process your ascii point observations and derive
those
variables yourself.
- Or, assuming you're running UPP, edit the configuration to dump out
forecast variables which match the type of observations you already
have.
Sorry I don't have a better answer for you. If you decide to do the
former, I can point you to the place in the MET code that does these
derivations, and you could copy the logic.
Thanks,
John Halley Gotway
met_help at ucar.edu
On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT
<met_help at ucar.edu>
wrote:
>
> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
> Queue: met_help
> Subject: ascii2nc followup
> Owner: Nobody
> Requestors: Mariusz.Pagowski at noaa.gov
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>
>
> Hello,
>
> I have observations of dew point and 6-hr accumulated precipitation,
write
> them out to ascii file setting codes 17 and 61, respectively, and
convert
> to prepbufer with ascii2nc.
>
> I later want to use pointstat* for WRF evaluation against these
data.
> There are no corresponding variables in WRF output. Do I need to
explicitly
> calculate
> mixing ratio from Td and some other form of precip from these
observations
> so they correspond to WRF output variables? Or there is some trick
that is
> available in pointstat configuration that would do it for me?
>
> Thanks,
> Mariusz
>
>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #71614] ascii2nc followup
From: Mariusz Pagowski
Time: Fri Apr 24 10:43:44 2015
John,
thanks, yes, I would like to use MET code for calculating moisture
related variables. I have QV output from wrf so it would be easiest to
convert
observations of Td to QV. Because of WRF file sizes I would rather
work with processing observations than WRF. I can derive 6hr
accumulated precip
from WRF with nco but that is a hog. If you have a better way of
dealing with it please point me to this solution.
Thanks,
Mariusz
> On Apr 24, 2015, at 10:12 AM, John Halley Gotway via RT
<met_help at ucar.edu> wrote:
>
> Mariusz,
>
> While the PB2NC tool has the ability to derive observation types
from input
> point observations (i.e. convert specific humidity and temperature
to
> relative humidity), the ASCII2NC and MADIS2NC tools do not. This
has come
> up in the past, and we would like to add this functionality but
haven't
> found the time or funding to do so yet.
>
> Your choices are...
> - Write a script to process your ascii point observations and derive
those
> variables yourself.
> - Or, assuming you're running UPP, edit the configuration to dump
out
> forecast variables which match the type of observations you already
have.
>
> Sorry I don't have a better answer for you. If you decide to do the
> former, I can point you to the place in the MET code that does these
> derivations, and you could copy the logic.
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
>
>
> On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT
<met_help at ucar.edu>
> wrote:
>
>>
>> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>> Queue: met_help
>> Subject: ascii2nc followup
>> Owner: Nobody
>> Requestors: Mariusz.Pagowski at noaa.gov
>> Status: new
>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>
>>
>> Hello,
>>
>> I have observations of dew point and 6-hr accumulated
precipitation, write
>> them out to ascii file setting codes 17 and 61, respectively, and
convert
>> to prepbufer with ascii2nc.
>>
>> I later want to use pointstat* for WRF evaluation against these
data.
>> There are no corresponding variables in WRF output. Do I need to
explicitly
>> calculate
>> mixing ratio from Td and some other form of precip from these
observations
>> so they correspond to WRF output variables? Or there is some trick
that is
>> available in pointstat configuration that would do it for me?
>>
>> Thanks,
>> Mariusz
>>
>>
>>
>>
>>
>
------------------------------------------------
Subject: ascii2nc followup
From: John Halley Gotway
Time: Fri Apr 24 10:56:14 2015
Mariusz,
The PB2NC tool has the ability to derive mean sea level pressure,
mixing
ratio, dewpoint temperature, relative humidity, wind direction, and
wind
speed.
This is done in the function named " derive_grib_code()" on line 1314
of
this file:
met-5.0/src/tools/other/pb2nc/pb2nc.cc
That function calls several "convert" routines that are define in the
MET
library code here:
met-5.0/src/basic/vx_util/conversions.cc
Hope that helps.
John
On Fri, Apr 24, 2015 at 10:43 AM, Mariusz Pagowski via RT
<met_help at ucar.edu
> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>
> John,
> thanks, yes, I would like to use MET code for calculating moisture
related
> variables. I have QV output from wrf so it would be easiest to
convert
> observations of Td to QV. Because of WRF file sizes I would rather
work
> with processing observations than WRF. I can derive 6hr accumulated
precip
> from WRF with nco but that is a hog. If you have a better way of
dealing
> with it please point me to this solution.
> Thanks,
> Mariusz
>
>
>
> > On Apr 24, 2015, at 10:12 AM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
> >
> > Mariusz,
> >
> > While the PB2NC tool has the ability to derive observation types
from
> input
> > point observations (i.e. convert specific humidity and temperature
to
> > relative humidity), the ASCII2NC and MADIS2NC tools do not. This
has
> come
> > up in the past, and we would like to add this functionality but
haven't
> > found the time or funding to do so yet.
> >
> > Your choices are...
> > - Write a script to process your ascii point observations and
derive
> those
> > variables yourself.
> > - Or, assuming you're running UPP, edit the configuration to dump
out
> > forecast variables which match the type of observations you
already have.
> >
> > Sorry I don't have a better answer for you. If you decide to do
the
> > former, I can point you to the place in the MET code that does
these
> > derivations, and you could copy the logic.
> >
> > Thanks,
> > John Halley Gotway
> > met_help at ucar.edu
> >
> >
> >
> > On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT <
> met_help at ucar.edu>
> > wrote:
> >
> >>
> >> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
> >> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
> >> Queue: met_help
> >> Subject: ascii2nc followup
> >> Owner: Nobody
> >> Requestors: Mariusz.Pagowski at noaa.gov
> >> Status: new
> >> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
> >>
> >>
> >> Hello,
> >>
> >> I have observations of dew point and 6-hr accumulated
precipitation,
> write
> >> them out to ascii file setting codes 17 and 61, respectively, and
> convert
> >> to prepbufer with ascii2nc.
> >>
> >> I later want to use pointstat* for WRF evaluation against these
data.
> >> There are no corresponding variables in WRF output. Do I need to
> explicitly
> >> calculate
> >> mixing ratio from Td and some other form of precip from these
> observations
> >> so they correspond to WRF output variables? Or there is some
trick that
> is
> >> available in pointstat configuration that would do it for me?
> >>
> >> Thanks,
> >> Mariusz
> >>
> >>
> >>
> >>
> >>
> >
>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #71614] ascii2nc followup
From: Mariusz Pagowski
Time: Thu Apr 30 15:26:28 2015
John,
there is one more question regarding wind.
In my obs file obtained by asccii2nc I have codes 31 and 32 for wind
direction and speed. In WRF file from pinterp I have variables UU/ VV
that I used to match with UGRD/VGRD when using your software for
converting prepbufr to nc. UU/VV are in “relative" WRF coordinates but
31/32 are “global” (zonal/meridional)
It seems I need to do something with 31 and 32 to match UU/VV. Can you
refer me where I can find these conversions in your codes? I assume
poinstat
will take care of UU/VV in WRF “relative” coords and convert them to
“global” so that statistics in output files are for zonal/meridional
winds - is that right?
Thanks,
Mariusz
> On Apr 24, 2015, at 10:56 AM, John Halley Gotway via RT
<met_help at ucar.edu> wrote:
>
> Mariusz,
>
> The PB2NC tool has the ability to derive mean sea level pressure,
mixing
> ratio, dewpoint temperature, relative humidity, wind direction, and
wind
> speed.
>
> This is done in the function named " derive_grib_code()" on line
1314 of
> this file:
> met-5.0/src/tools/other/pb2nc/pb2nc.cc
>
> That function calls several "convert" routines that are define in
the MET
> library code here:
> met-5.0/src/basic/vx_util/conversions.cc
>
> Hope that helps.
>
> John
>
> On Fri, Apr 24, 2015 at 10:43 AM, Mariusz Pagowski via RT
<met_help at ucar.edu
>> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>
>> John,
>> thanks, yes, I would like to use MET code for calculating moisture
related
>> variables. I have QV output from wrf so it would be easiest to
convert
>> observations of Td to QV. Because of WRF file sizes I would rather
work
>> with processing observations than WRF. I can derive 6hr accumulated
precip
>> from WRF with nco but that is a hog. If you have a better way of
dealing
>> with it please point me to this solution.
>> Thanks,
>> Mariusz
>>
>>
>>
>>> On Apr 24, 2015, at 10:12 AM, John Halley Gotway via RT <
>> met_help at ucar.edu> wrote:
>>>
>>> Mariusz,
>>>
>>> While the PB2NC tool has the ability to derive observation types
from
>> input
>>> point observations (i.e. convert specific humidity and temperature
to
>>> relative humidity), the ASCII2NC and MADIS2NC tools do not. This
has
>> come
>>> up in the past, and we would like to add this functionality but
haven't
>>> found the time or funding to do so yet.
>>>
>>> Your choices are...
>>> - Write a script to process your ascii point observations and
derive
>> those
>>> variables yourself.
>>> - Or, assuming you're running UPP, edit the configuration to dump
out
>>> forecast variables which match the type of observations you
already have.
>>>
>>> Sorry I don't have a better answer for you. If you decide to do
the
>>> former, I can point you to the place in the MET code that does
these
>>> derivations, and you could copy the logic.
>>>
>>> Thanks,
>>> John Halley Gotway
>>> met_help at ucar.edu
>>>
>>>
>>>
>>> On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT <
>> met_help at ucar.edu>
>>> wrote:
>>>
>>>>
>>>> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>> Queue: met_help
>>>> Subject: ascii2nc followup
>>>> Owner: Nobody
>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>> Status: new
>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>>>
>>>>
>>>> Hello,
>>>>
>>>> I have observations of dew point and 6-hr accumulated
precipitation,
>> write
>>>> them out to ascii file setting codes 17 and 61, respectively, and
>> convert
>>>> to prepbufer with ascii2nc.
>>>>
>>>> I later want to use pointstat* for WRF evaluation against these
data.
>>>> There are no corresponding variables in WRF output. Do I need to
>> explicitly
>>>> calculate
>>>> mixing ratio from Td and some other form of precip from these
>> observations
>>>> so they correspond to WRF output variables? Or there is some
trick that
>> is
>>>> available in pointstat configuration that would do it for me?
>>>>
>>>> Thanks,
>>>> Mariusz
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>>
>>
>
------------------------------------------------
Subject: ascii2nc followup
From: John Halley Gotway
Time: Fri May 01 13:24:58 2015
Mariusz,
I see that you're asking about rotating winds from being grid-relative
to
being earth-relative.
I assume you're using UPP to post-process your WRF files and have
output in
GRIB format. GRIB includes a header value to indicate whether winds
are
defined as earth or grid-relative. The MET Point-Stat tool assumes
that
your point observations are defined as earth-relative (as they should
be).
It checks the GRIB header and automatically rotates them from grid-to-
earth
relative for you.
If you run Point-Stat at verbosity level 3, you'll see debug messages
similar to the following indicating that this rotation is occurring:
DEBUG 2: Reading data for UGRD/Z10.
DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found range match
for
VarInfo "UGRD/Z10" in GRIB record 37 of GRIB file
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
DEBUG 3: MetGrib1DataFile::rotate_winds() -> Have U-wind record,
reading
V-wind record.
DEBUG 3: MetGrib1DataFile::data_plane_scalar() -> Found exact match
for
VarInfo "UGRD/Z10" in GRIB record 38 of GRIB file
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
DEBUG 3: Rotating U and V wind components from grid-relative to
earth-relative.
DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found 1 GRIB
records
matching VarInfo "UGRD/Z10" in GRIB file
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
DEBUG 2: For UGRD/Z10 found 1 forecast levels and 0 climatology
levels.
I the example above, we're verifying the U-component of wind, UGRD.
In
order to rotate from grid to earth relative, it reads in VGRD and does
the
rotation. As long as you're using GRIB files, that rotation should
happen
automatically for you.
Additionally, if you select WIND (wind speed) to be verified, Point-
Stat
looks in the input GRIB file and uses WIND if it's present. If it's
not
present, it'll automatically read in UGRD and VGRD and derive it for
you.
Point-Stat can be used to evaluation UGRD (U-wind component), VGRD (V-
wind
component), and WIND (wind speed). However, it doesn't evaluate wind
direction directly. Since wind direction is a "circular" variable, it
can't be treated the same as the other "scalar" variables. My
suggestions
is this... run Point-Stat to evaluate UGRD, VGRD, and WIND. And be
sure to
write out the VL1L2 output line. You could also write the matched
pair
(MPR) output line. Then you can run the VL1L2 and/or MPR output lines
through the STAT-Analysis "WDIR" job to evaluate wind direction.
Thanks,
John
On Thu, Apr 30, 2015 at 3:26 PM, Mariusz Pagowski via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>
> John,
> there is one more question regarding wind.
> In my obs file obtained by asccii2nc I have codes 31 and 32 for wind
> direction and speed. In WRF file from pinterp I have variables UU/
VV that
> I used to match with UGRD/VGRD when using your software for
converting
> prepbufr to nc. UU/VV are in “relative" WRF coordinates but 31/32
are
> “global” (zonal/meridional)
>
> It seems I need to do something with 31 and 32 to match UU/VV. Can
you
> refer me where I can find these conversions in your codes? I assume
poinstat
> will take care of UU/VV in WRF “relative” coords and convert them
to
> “global” so that statistics in output files are for zonal/meridional
winds
> - is that right?
>
> Thanks,
> Mariusz
>
>
>
>
> > On Apr 24, 2015, at 10:56 AM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
> >
> > Mariusz,
> >
> > The PB2NC tool has the ability to derive mean sea level pressure,
mixing
> > ratio, dewpoint temperature, relative humidity, wind direction,
and wind
> > speed.
> >
> > This is done in the function named " derive_grib_code()" on line
1314 of
> > this file:
> > met-5.0/src/tools/other/pb2nc/pb2nc.cc
> >
> > That function calls several "convert" routines that are define in
the MET
> > library code here:
> > met-5.0/src/basic/vx_util/conversions.cc
> >
> > Hope that helps.
> >
> > John
> >
> > On Fri, Apr 24, 2015 at 10:43 AM, Mariusz Pagowski via RT <
> met_help at ucar.edu
> >> wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
> >>
> >> John,
> >> thanks, yes, I would like to use MET code for calculating
moisture
> related
> >> variables. I have QV output from wrf so it would be easiest to
convert
> >> observations of Td to QV. Because of WRF file sizes I would
rather work
> >> with processing observations than WRF. I can derive 6hr
accumulated
> precip
> >> from WRF with nco but that is a hog. If you have a better way of
dealing
> >> with it please point me to this solution.
> >> Thanks,
> >> Mariusz
> >>
> >>
> >>
> >>> On Apr 24, 2015, at 10:12 AM, John Halley Gotway via RT <
> >> met_help at ucar.edu> wrote:
> >>>
> >>> Mariusz,
> >>>
> >>> While the PB2NC tool has the ability to derive observation types
from
> >> input
> >>> point observations (i.e. convert specific humidity and
temperature to
> >>> relative humidity), the ASCII2NC and MADIS2NC tools do not.
This has
> >> come
> >>> up in the past, and we would like to add this functionality but
haven't
> >>> found the time or funding to do so yet.
> >>>
> >>> Your choices are...
> >>> - Write a script to process your ascii point observations and
derive
> >> those
> >>> variables yourself.
> >>> - Or, assuming you're running UPP, edit the configuration to
dump out
> >>> forecast variables which match the type of observations you
already
> have.
> >>>
> >>> Sorry I don't have a better answer for you. If you decide to do
the
> >>> former, I can point you to the place in the MET code that does
these
> >>> derivations, and you could copy the logic.
> >>>
> >>> Thanks,
> >>> John Halley Gotway
> >>> met_help at ucar.edu
> >>>
> >>>
> >>>
> >>> On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT <
> >> met_help at ucar.edu>
> >>> wrote:
> >>>
> >>>>
> >>>> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
> >>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
> >>>> Queue: met_help
> >>>> Subject: ascii2nc followup
> >>>> Owner: Nobody
> >>>> Requestors: Mariusz.Pagowski at noaa.gov
> >>>> Status: new
> >>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614
> >
> >>>>
> >>>>
> >>>> Hello,
> >>>>
> >>>> I have observations of dew point and 6-hr accumulated
precipitation,
> >> write
> >>>> them out to ascii file setting codes 17 and 61, respectively,
and
> >> convert
> >>>> to prepbufer with ascii2nc.
> >>>>
> >>>> I later want to use pointstat* for WRF evaluation against these
data.
> >>>> There are no corresponding variables in WRF output. Do I need
to
> >> explicitly
> >>>> calculate
> >>>> mixing ratio from Td and some other form of precip from these
> >> observations
> >>>> so they correspond to WRF output variables? Or there is some
trick
> that
> >> is
> >>>> available in pointstat configuration that would do it for me?
> >>>>
> >>>> Thanks,
> >>>> Mariusz
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>
> >>
> >>
> >>
> >
>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #71614] ascii2nc followup
From: Mariusz Pagowski
Time: Fri May 01 13:46:06 2015
John,
there is one important issue. I don’t use UPP - I post process WRF
with pinterp so that UU, VV are not rotated to earth-relative. Then I
need to rotate the winds somewhere. I wondered if point-stat will do
it for me, but looks like it won’t. Can you confirm. Then I need to
modify interp to output earth-relative wind to have correct wind
evaluation with point-stat - right?
Mariusz
> On May 1, 2015, at 1:24 PM, John Halley Gotway via RT
<met_help at ucar.edu> wrote:
>
> Mariusz,
>
> I see that you're asking about rotating winds from being grid-
relative to
> being earth-relative.
>
> I assume you're using UPP to post-process your WRF files and have
output in
> GRIB format. GRIB includes a header value to indicate whether winds
are
> defined as earth or grid-relative. The MET Point-Stat tool assumes
that
> your point observations are defined as earth-relative (as they
should be).
> It checks the GRIB header and automatically rotates them from grid-
to-earth
> relative for you.
>
> If you run Point-Stat at verbosity level 3, you'll see debug
messages
> similar to the following indicating that this rotation is occurring:
>
> DEBUG 2: Reading data for UGRD/Z10.
> DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found range match
for
> VarInfo "UGRD/Z10" in GRIB record 37 of GRIB file
>
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
> DEBUG 3: MetGrib1DataFile::rotate_winds() -> Have U-wind record,
reading
> V-wind record.
> DEBUG 3: MetGrib1DataFile::data_plane_scalar() -> Found exact
match for
> VarInfo "UGRD/Z10" in GRIB record 38 of GRIB file
>
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
> DEBUG 3: Rotating U and V wind components from grid-relative to
> earth-relative.
> DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found 1 GRIB
records
> matching VarInfo "UGRD/Z10" in GRIB file
>
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
> DEBUG 2: For UGRD/Z10 found 1 forecast levels and 0 climatology
levels.
>
> I the example above, we're verifying the U-component of wind, UGRD.
In
> order to rotate from grid to earth relative, it reads in VGRD and
does the
> rotation. As long as you're using GRIB files, that rotation should
happen
> automatically for you.
>
> Additionally, if you select WIND (wind speed) to be verified, Point-
Stat
> looks in the input GRIB file and uses WIND if it's present. If it's
not
> present, it'll automatically read in UGRD and VGRD and derive it for
you.
>
> Point-Stat can be used to evaluation UGRD (U-wind component), VGRD
(V-wind
> component), and WIND (wind speed). However, it doesn't evaluate
wind
> direction directly. Since wind direction is a "circular" variable,
it
> can't be treated the same as the other "scalar" variables. My
suggestions
> is this... run Point-Stat to evaluate UGRD, VGRD, and WIND. And be
sure to
> write out the VL1L2 output line. You could also write the matched
pair
> (MPR) output line. Then you can run the VL1L2 and/or MPR output
lines
> through the STAT-Analysis "WDIR" job to evaluate wind direction.
>
> Thanks,
> John
>
> On Thu, Apr 30, 2015 at 3:26 PM, Mariusz Pagowski via RT
<met_help at ucar.edu>
> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>
>> John,
>> there is one more question regarding wind.
>> In my obs file obtained by asccii2nc I have codes 31 and 32 for
wind
>> direction and speed. In WRF file from pinterp I have variables UU/
VV that
>> I used to match with UGRD/VGRD when using your software for
converting
>> prepbufr to nc. UU/VV are in “relative" WRF coordinates but 31/32
are
>> “global” (zonal/meridional)
>>
>> It seems I need to do something with 31 and 32 to match UU/VV. Can
you
>> refer me where I can find these conversions in your codes? I assume
poinstat
>> will take care of UU/VV in WRF “relative” coords and convert them
to
>> “global” so that statistics in output files are for
zonal/meridional winds
>> - is that right?
>>
>> Thanks,
>> Mariusz
>>
>>
>>
>>
>>> On Apr 24, 2015, at 10:56 AM, John Halley Gotway via RT <
>> met_help at ucar.edu> wrote:
>>>
>>> Mariusz,
>>>
>>> The PB2NC tool has the ability to derive mean sea level pressure,
mixing
>>> ratio, dewpoint temperature, relative humidity, wind direction,
and wind
>>> speed.
>>>
>>> This is done in the function named " derive_grib_code()" on line
1314 of
>>> this file:
>>> met-5.0/src/tools/other/pb2nc/pb2nc.cc
>>>
>>> That function calls several "convert" routines that are define in
the MET
>>> library code here:
>>> met-5.0/src/basic/vx_util/conversions.cc
>>>
>>> Hope that helps.
>>>
>>> John
>>>
>>> On Fri, Apr 24, 2015 at 10:43 AM, Mariusz Pagowski via RT <
>> met_help at ucar.edu
>>>> wrote:
>>>
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>>>
>>>> John,
>>>> thanks, yes, I would like to use MET code for calculating
moisture
>> related
>>>> variables. I have QV output from wrf so it would be easiest to
convert
>>>> observations of Td to QV. Because of WRF file sizes I would
rather work
>>>> with processing observations than WRF. I can derive 6hr
accumulated
>> precip
>>>> from WRF with nco but that is a hog. If you have a better way of
dealing
>>>> with it please point me to this solution.
>>>> Thanks,
>>>> Mariusz
>>>>
>>>>
>>>>
>>>>> On Apr 24, 2015, at 10:12 AM, John Halley Gotway via RT <
>>>> met_help at ucar.edu> wrote:
>>>>>
>>>>> Mariusz,
>>>>>
>>>>> While the PB2NC tool has the ability to derive observation types
from
>>>> input
>>>>> point observations (i.e. convert specific humidity and
temperature to
>>>>> relative humidity), the ASCII2NC and MADIS2NC tools do not.
This has
>>>> come
>>>>> up in the past, and we would like to add this functionality but
haven't
>>>>> found the time or funding to do so yet.
>>>>>
>>>>> Your choices are...
>>>>> - Write a script to process your ascii point observations and
derive
>>>> those
>>>>> variables yourself.
>>>>> - Or, assuming you're running UPP, edit the configuration to
dump out
>>>>> forecast variables which match the type of observations you
already
>> have.
>>>>>
>>>>> Sorry I don't have a better answer for you. If you decide to do
the
>>>>> former, I can point you to the place in the MET code that does
these
>>>>> derivations, and you could copy the logic.
>>>>>
>>>>> Thanks,
>>>>> John Halley Gotway
>>>>> met_help at ucar.edu
>>>>>
>>>>>
>>>>>
>>>>> On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT <
>>>> met_help at ucar.edu>
>>>>> wrote:
>>>>>
>>>>>>
>>>>>> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
>>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>>> Queue: met_help
>>>>>> Subject: ascii2nc followup
>>>>>> Owner: Nobody
>>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>>> Status: new
>>>>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614
>>>
>>>>>>
>>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> I have observations of dew point and 6-hr accumulated
precipitation,
>>>> write
>>>>>> them out to ascii file setting codes 17 and 61, respectively,
and
>>>> convert
>>>>>> to prepbufer with ascii2nc.
>>>>>>
>>>>>> I later want to use pointstat* for WRF evaluation against these
data.
>>>>>> There are no corresponding variables in WRF output. Do I need
to
>>>> explicitly
>>>>>> calculate
>>>>>> mixing ratio from Td and some other form of precip from these
>>>> observations
>>>>>> so they correspond to WRF output variables? Or there is some
trick
>> that
>>>> is
>>>>>> available in pointstat configuration that would do it for me?
>>>>>>
>>>>>> Thanks,
>>>>>> Mariusz
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>>
>>
>
------------------------------------------------
Subject: ascii2nc followup
From: John Halley Gotway
Time: Fri May 01 14:06:45 2015
Mariusz,
One huge drawback of using the WRF pinterp utility is that it can't
read
variables that are defined on the staggered dimensions. Winds are
defined
on the staggered dimensions, and therefore can't be evaluated in MET
using
pinterp data.
Here's an example of trying to read UGRD from a pinterp file with the
MET
plot_data_plane utility:
plot_data_plane wrfout_d01_2008-08-08_12:00:00_PLEV uu.ps 'name="UU";
level="(0,0,*,*)";'
DEBUG 1: Opening data file:
/var/autofs/mnt/dakota_d3/projects/MET/MET_test_data/unit_test/model_data/p_interp/wrfout_d01_2008-
08-08_12:00:00_PLEV
ERROR :
ERROR : PinterpFile::data(NcVar *, const LongArray &, DataPlane &,
double
&) const -> can't find needed dimensions(s) for variable "UU" ... one
of
the dimensions may be staggered.
ERROR :
MET can't be used to evaluate winds from pinterp files. This is
mentioned
in the MET online tutorial on this page:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/rec_software.php
I'd suggest using UPP instead.
Thanks,
John
On Fri, May 1, 2015 at 1:46 PM, Mariusz Pagowski via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>
> John,
> there is one important issue. I don’t use UPP - I post process WRF
with
> pinterp so that UU, VV are not rotated to earth-relative. Then I
need to
> rotate the winds somewhere. I wondered if point-stat will do it for
me, but
> looks like it won’t. Can you confirm. Then I need to modify interp
to
> output earth-relative wind to have correct wind evaluation with
point-stat
> - right?
> Mariusz
>
>
>
> > On May 1, 2015, at 1:24 PM, John Halley Gotway via RT
<met_help at ucar.edu>
> wrote:
> >
> > Mariusz,
> >
> > I see that you're asking about rotating winds from being grid-
relative to
> > being earth-relative.
> >
> > I assume you're using UPP to post-process your WRF files and have
output
> in
> > GRIB format. GRIB includes a header value to indicate whether
winds are
> > defined as earth or grid-relative. The MET Point-Stat tool
assumes that
> > your point observations are defined as earth-relative (as they
should
> be).
> > It checks the GRIB header and automatically rotates them from
> grid-to-earth
> > relative for you.
> >
> > If you run Point-Stat at verbosity level 3, you'll see debug
messages
> > similar to the following indicating that this rotation is
occurring:
> >
> > DEBUG 2: Reading data for UGRD/Z10.
> > DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found range
match for
> > VarInfo "UGRD/Z10" in GRIB record 37 of GRIB file
> >
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
> > DEBUG 3: MetGrib1DataFile::rotate_winds() -> Have U-wind record,
> reading
> > V-wind record.
> > DEBUG 3: MetGrib1DataFile::data_plane_scalar() -> Found exact
match for
> > VarInfo "UGRD/Z10" in GRIB record 38 of GRIB file
> >
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
> > DEBUG 3: Rotating U and V wind components from grid-relative to
> > earth-relative.
> > DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found 1 GRIB
records
> > matching VarInfo "UGRD/Z10" in GRIB file
> >
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
> > DEBUG 2: For UGRD/Z10 found 1 forecast levels and 0 climatology
levels.
> >
> > I the example above, we're verifying the U-component of wind,
UGRD. In
> > order to rotate from grid to earth relative, it reads in VGRD and
does
> the
> > rotation. As long as you're using GRIB files, that rotation
should
> happen
> > automatically for you.
> >
> > Additionally, if you select WIND (wind speed) to be verified,
Point-Stat
> > looks in the input GRIB file and uses WIND if it's present. If
it's not
> > present, it'll automatically read in UGRD and VGRD and derive it
for you.
> >
> > Point-Stat can be used to evaluation UGRD (U-wind component), VGRD
> (V-wind
> > component), and WIND (wind speed). However, it doesn't evaluate
wind
> > direction directly. Since wind direction is a "circular"
variable, it
> > can't be treated the same as the other "scalar" variables. My
> suggestions
> > is this... run Point-Stat to evaluate UGRD, VGRD, and WIND. And
be sure
> to
> > write out the VL1L2 output line. You could also write the matched
pair
> > (MPR) output line. Then you can run the VL1L2 and/or MPR output
lines
> > through the STAT-Analysis "WDIR" job to evaluate wind direction.
> >
> > Thanks,
> > John
> >
> > On Thu, Apr 30, 2015 at 3:26 PM, Mariusz Pagowski via RT <
> met_help at ucar.edu>
> > wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
> >>
> >> John,
> >> there is one more question regarding wind.
> >> In my obs file obtained by asccii2nc I have codes 31 and 32 for
wind
> >> direction and speed. In WRF file from pinterp I have variables
UU/ VV
> that
> >> I used to match with UGRD/VGRD when using your software for
converting
> >> prepbufr to nc. UU/VV are in “relative" WRF coordinates but 31/32
are
> >> “global” (zonal/meridional)
> >>
> >> It seems I need to do something with 31 and 32 to match UU/VV.
Can you
> >> refer me where I can find these conversions in your codes? I
assume
> poinstat
> >> will take care of UU/VV in WRF “relative” coords and convert
them to
> >> “global” so that statistics in output files are for
zonal/meridional
> winds
> >> - is that right?
> >>
> >> Thanks,
> >> Mariusz
> >>
> >>
> >>
> >>
> >>> On Apr 24, 2015, at 10:56 AM, John Halley Gotway via RT <
> >> met_help at ucar.edu> wrote:
> >>>
> >>> Mariusz,
> >>>
> >>> The PB2NC tool has the ability to derive mean sea level
pressure,
> mixing
> >>> ratio, dewpoint temperature, relative humidity, wind direction,
and
> wind
> >>> speed.
> >>>
> >>> This is done in the function named " derive_grib_code()" on line
1314
> of
> >>> this file:
> >>> met-5.0/src/tools/other/pb2nc/pb2nc.cc
> >>>
> >>> That function calls several "convert" routines that are define
in the
> MET
> >>> library code here:
> >>> met-5.0/src/basic/vx_util/conversions.cc
> >>>
> >>> Hope that helps.
> >>>
> >>> John
> >>>
> >>> On Fri, Apr 24, 2015 at 10:43 AM, Mariusz Pagowski via RT <
> >> met_help at ucar.edu
> >>>> wrote:
> >>>
> >>>>
> >>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
> >>>>
> >>>> John,
> >>>> thanks, yes, I would like to use MET code for calculating
moisture
> >> related
> >>>> variables. I have QV output from wrf so it would be easiest to
convert
> >>>> observations of Td to QV. Because of WRF file sizes I would
rather
> work
> >>>> with processing observations than WRF. I can derive 6hr
accumulated
> >> precip
> >>>> from WRF with nco but that is a hog. If you have a better way
of
> dealing
> >>>> with it please point me to this solution.
> >>>> Thanks,
> >>>> Mariusz
> >>>>
> >>>>
> >>>>
> >>>>> On Apr 24, 2015, at 10:12 AM, John Halley Gotway via RT <
> >>>> met_help at ucar.edu> wrote:
> >>>>>
> >>>>> Mariusz,
> >>>>>
> >>>>> While the PB2NC tool has the ability to derive observation
types from
> >>>> input
> >>>>> point observations (i.e. convert specific humidity and
temperature to
> >>>>> relative humidity), the ASCII2NC and MADIS2NC tools do not.
This has
> >>>> come
> >>>>> up in the past, and we would like to add this functionality
but
> haven't
> >>>>> found the time or funding to do so yet.
> >>>>>
> >>>>> Your choices are...
> >>>>> - Write a script to process your ascii point observations and
derive
> >>>> those
> >>>>> variables yourself.
> >>>>> - Or, assuming you're running UPP, edit the configuration to
dump out
> >>>>> forecast variables which match the type of observations you
already
> >> have.
> >>>>>
> >>>>> Sorry I don't have a better answer for you. If you decide to
do the
> >>>>> former, I can point you to the place in the MET code that does
these
> >>>>> derivations, and you could copy the logic.
> >>>>>
> >>>>> Thanks,
> >>>>> John Halley Gotway
> >>>>> met_help at ucar.edu
> >>>>>
> >>>>>
> >>>>>
> >>>>> On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT <
> >>>> met_help at ucar.edu>
> >>>>> wrote:
> >>>>>
> >>>>>>
> >>>>>> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
> >>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
> >>>>>> Queue: met_help
> >>>>>> Subject: ascii2nc followup
> >>>>>> Owner: Nobody
> >>>>>> Requestors: Mariusz.Pagowski at noaa.gov
> >>>>>> Status: new
> >>>>>> Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614
> >>>
> >>>>>>
> >>>>>>
> >>>>>> Hello,
> >>>>>>
> >>>>>> I have observations of dew point and 6-hr accumulated
precipitation,
> >>>> write
> >>>>>> them out to ascii file setting codes 17 and 61, respectively,
and
> >>>> convert
> >>>>>> to prepbufer with ascii2nc.
> >>>>>>
> >>>>>> I later want to use pointstat* for WRF evaluation against
these
> data.
> >>>>>> There are no corresponding variables in WRF output. Do I need
to
> >>>> explicitly
> >>>>>> calculate
> >>>>>> mixing ratio from Td and some other form of precip from these
> >>>> observations
> >>>>>> so they correspond to WRF output variables? Or there is some
trick
> >> that
> >>>> is
> >>>>>> available in pointstat configuration that would do it for me?
> >>>>>>
> >>>>>> Thanks,
> >>>>>> Mariusz
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>
> >>
> >>
> >>
> >
>
>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #71614] ascii2nc followup
From: Mariusz Pagowski
Time: Fri May 01 14:19:16 2015
John,
pinterp will unstagger wind if you set option for that. Anyway, I see
you guys are switching to UPP. I have been using MET 4.1 so far.
Before I go for UPP and grib, that will change my whole post
processing, I need to rotate the wind in pinterp.
Thanks for your help,
Mariusz
> On May 1, 2015, at 2:06 PM, John Halley Gotway via RT
<met_help at ucar.edu> wrote:
>
> Mariusz,
>
> One huge drawback of using the WRF pinterp utility is that it can't
read
> variables that are defined on the staggered dimensions. Winds are
defined
> on the staggered dimensions, and therefore can't be evaluated in MET
using
> pinterp data.
>
> Here's an example of trying to read UGRD from a pinterp file with
the MET
> plot_data_plane utility:
>
> plot_data_plane wrfout_d01_2008-08-08_12:00:00_PLEV uu.ps
'name="UU";
> level="(0,0,*,*)";'
>
> DEBUG 1: Opening data file:
>
/var/autofs/mnt/dakota_d3/projects/MET/MET_test_data/unit_test/model_data/p_interp/wrfout_d01_2008-
08-08_12:00:00_PLEV
> ERROR :
> ERROR : PinterpFile::data(NcVar *, const LongArray &, DataPlane &,
double
> &) const -> can't find needed dimensions(s) for variable "UU" ...
one of
> the dimensions may be staggered.
> ERROR :
>
> MET can't be used to evaluate winds from pinterp files. This is
mentioned
> in the MET online tutorial on this page:
>
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/compilation/rec_software.php
>
> I'd suggest using UPP instead.
>
> Thanks,
> John
>
>
> On Fri, May 1, 2015 at 1:46 PM, Mariusz Pagowski via RT
<met_help at ucar.edu>
> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>
>> John,
>> there is one important issue. I don’t use UPP - I post process WRF
with
>> pinterp so that UU, VV are not rotated to earth-relative. Then I
need to
>> rotate the winds somewhere. I wondered if point-stat will do it for
me, but
>> looks like it won’t. Can you confirm. Then I need to modify interp
to
>> output earth-relative wind to have correct wind evaluation with
point-stat
>> - right?
>> Mariusz
>>
>>
>>
>>> On May 1, 2015, at 1:24 PM, John Halley Gotway via RT
<met_help at ucar.edu>
>> wrote:
>>>
>>> Mariusz,
>>>
>>> I see that you're asking about rotating winds from being grid-
relative to
>>> being earth-relative.
>>>
>>> I assume you're using UPP to post-process your WRF files and have
output
>> in
>>> GRIB format. GRIB includes a header value to indicate whether
winds are
>>> defined as earth or grid-relative. The MET Point-Stat tool
assumes that
>>> your point observations are defined as earth-relative (as they
should
>> be).
>>> It checks the GRIB header and automatically rotates them from
>> grid-to-earth
>>> relative for you.
>>>
>>> If you run Point-Stat at verbosity level 3, you'll see debug
messages
>>> similar to the following indicating that this rotation is
occurring:
>>>
>>> DEBUG 2: Reading data for UGRD/Z10.
>>> DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found range
match for
>>> VarInfo "UGRD/Z10" in GRIB record 37 of GRIB file
>>>
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
>>> DEBUG 3: MetGrib1DataFile::rotate_winds() -> Have U-wind record,
>> reading
>>> V-wind record.
>>> DEBUG 3: MetGrib1DataFile::data_plane_scalar() -> Found exact
match for
>>> VarInfo "UGRD/Z10" in GRIB record 38 of GRIB file
>>>
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
>>> DEBUG 3: Rotating U and V wind components from grid-relative to
>>> earth-relative.
>>> DEBUG 3: MetGrib1DataFile::data_plane_array() -> Found 1 GRIB
records
>>> matching VarInfo "UGRD/Z10" in GRIB file
>>>
"../data/sample_fcst/2007033000/nam.t00z.awip1236.tm00.20070330.grb".
>>> DEBUG 2: For UGRD/Z10 found 1 forecast levels and 0 climatology
levels.
>>>
>>> I the example above, we're verifying the U-component of wind,
UGRD. In
>>> order to rotate from grid to earth relative, it reads in VGRD and
does
>> the
>>> rotation. As long as you're using GRIB files, that rotation
should
>> happen
>>> automatically for you.
>>>
>>> Additionally, if you select WIND (wind speed) to be verified,
Point-Stat
>>> looks in the input GRIB file and uses WIND if it's present. If
it's not
>>> present, it'll automatically read in UGRD and VGRD and derive it
for you.
>>>
>>> Point-Stat can be used to evaluation UGRD (U-wind component), VGRD
>> (V-wind
>>> component), and WIND (wind speed). However, it doesn't evaluate
wind
>>> direction directly. Since wind direction is a "circular"
variable, it
>>> can't be treated the same as the other "scalar" variables. My
>> suggestions
>>> is this... run Point-Stat to evaluate UGRD, VGRD, and WIND. And
be sure
>> to
>>> write out the VL1L2 output line. You could also write the matched
pair
>>> (MPR) output line. Then you can run the VL1L2 and/or MPR output
lines
>>> through the STAT-Analysis "WDIR" job to evaluate wind direction.
>>>
>>> Thanks,
>>> John
>>>
>>> On Thu, Apr 30, 2015 at 3:26 PM, Mariusz Pagowski via RT <
>> met_help at ucar.edu>
>>> wrote:
>>>
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>>>
>>>> John,
>>>> there is one more question regarding wind.
>>>> In my obs file obtained by asccii2nc I have codes 31 and 32 for
wind
>>>> direction and speed. In WRF file from pinterp I have variables
UU/ VV
>> that
>>>> I used to match with UGRD/VGRD when using your software for
converting
>>>> prepbufr to nc. UU/VV are in “relative" WRF coordinates but 31/32
are
>>>> “global” (zonal/meridional)
>>>>
>>>> It seems I need to do something with 31 and 32 to match UU/VV.
Can you
>>>> refer me where I can find these conversions in your codes? I
assume
>> poinstat
>>>> will take care of UU/VV in WRF “relative” coords and convert
them to
>>>> “global” so that statistics in output files are for
zonal/meridional
>> winds
>>>> - is that right?
>>>>
>>>> Thanks,
>>>> Mariusz
>>>>
>>>>
>>>>
>>>>
>>>>> On Apr 24, 2015, at 10:56 AM, John Halley Gotway via RT <
>>>> met_help at ucar.edu> wrote:
>>>>>
>>>>> Mariusz,
>>>>>
>>>>> The PB2NC tool has the ability to derive mean sea level
pressure,
>> mixing
>>>>> ratio, dewpoint temperature, relative humidity, wind direction,
and
>> wind
>>>>> speed.
>>>>>
>>>>> This is done in the function named " derive_grib_code()" on line
1314
>> of
>>>>> this file:
>>>>> met-5.0/src/tools/other/pb2nc/pb2nc.cc
>>>>>
>>>>> That function calls several "convert" routines that are define
in the
>> MET
>>>>> library code here:
>>>>> met-5.0/src/basic/vx_util/conversions.cc
>>>>>
>>>>> Hope that helps.
>>>>>
>>>>> John
>>>>>
>>>>> On Fri, Apr 24, 2015 at 10:43 AM, Mariusz Pagowski via RT <
>>>> met_help at ucar.edu
>>>>>> wrote:
>>>>>
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614 >
>>>>>>
>>>>>> John,
>>>>>> thanks, yes, I would like to use MET code for calculating
moisture
>>>> related
>>>>>> variables. I have QV output from wrf so it would be easiest to
convert
>>>>>> observations of Td to QV. Because of WRF file sizes I would
rather
>> work
>>>>>> with processing observations than WRF. I can derive 6hr
accumulated
>>>> precip
>>>>>> from WRF with nco but that is a hog. If you have a better way
of
>> dealing
>>>>>> with it please point me to this solution.
>>>>>> Thanks,
>>>>>> Mariusz
>>>>>>
>>>>>>
>>>>>>
>>>>>>> On Apr 24, 2015, at 10:12 AM, John Halley Gotway via RT <
>>>>>> met_help at ucar.edu> wrote:
>>>>>>>
>>>>>>> Mariusz,
>>>>>>>
>>>>>>> While the PB2NC tool has the ability to derive observation
types from
>>>>>> input
>>>>>>> point observations (i.e. convert specific humidity and
temperature to
>>>>>>> relative humidity), the ASCII2NC and MADIS2NC tools do not.
This has
>>>>>> come
>>>>>>> up in the past, and we would like to add this functionality
but
>> haven't
>>>>>>> found the time or funding to do so yet.
>>>>>>>
>>>>>>> Your choices are...
>>>>>>> - Write a script to process your ascii point observations and
derive
>>>>>> those
>>>>>>> variables yourself.
>>>>>>> - Or, assuming you're running UPP, edit the configuration to
dump out
>>>>>>> forecast variables which match the type of observations you
already
>>>> have.
>>>>>>>
>>>>>>> Sorry I don't have a better answer for you. If you decide to
do the
>>>>>>> former, I can point you to the place in the MET code that does
these
>>>>>>> derivations, and you could copy the logic.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> John Halley Gotway
>>>>>>> met_help at ucar.edu
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On Thu, Apr 23, 2015 at 2:21 PM, Mariusz Pagowski via RT <
>>>>>> met_help at ucar.edu>
>>>>>>> wrote:
>>>>>>>
>>>>>>>>
>>>>>>>> Thu Apr 23 14:21:01 2015: Request 71614 was acted upon.
>>>>>>>> Transaction: Ticket created by Mariusz.Pagowski at noaa.gov
>>>>>>>> Queue: met_help
>>>>>>>> Subject: ascii2nc followup
>>>>>>>> Owner: Nobody
>>>>>>>> Requestors: Mariusz.Pagowski at noaa.gov
>>>>>>>> Status: new
>>>>>>>> Ticket <URL:
>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71614
>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Hello,
>>>>>>>>
>>>>>>>> I have observations of dew point and 6-hr accumulated
precipitation,
>>>>>> write
>>>>>>>> them out to ascii file setting codes 17 and 61, respectively,
and
>>>>>> convert
>>>>>>>> to prepbufer with ascii2nc.
>>>>>>>>
>>>>>>>> I later want to use pointstat* for WRF evaluation against
these
>> data.
>>>>>>>> There are no corresponding variables in WRF output. Do I need
to
>>>>>> explicitly
>>>>>>>> calculate
>>>>>>>> mixing ratio from Td and some other form of precip from these
>>>>>> observations
>>>>>>>> so they correspond to WRF output variables? Or there is some
trick
>>>> that
>>>>>> is
>>>>>>>> available in pointstat configuration that would do it for me?
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Mariusz
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>>
>>
>
------------------------------------------------
Subject: ascii2nc followup
From: John Halley Gotway
Time: Fri May 01 14:35:15 2015
Mariusz,
Hmmm, I wasn't aware that pinterp coiuld destagger winds. That's
great!
I've only run it enough to generate the output we use in our nightly
regression tests. Clearly, we need to generate a wider variety of
test
data.
Even if MET is able to read the wind fields, it doesn't currently have
logic to rotate them. I can see how we could port the GRIB1 wind
rotation
logic over for p_interp, but that'd be a development task. I could
add it
to our list.
This rotation is done in the file
met-5.0/src/libcode/vx_data2d/data2d_utils.cc in a function named
"rotate_uv_grid_to_earth()".
For now, I believe you're stuck with either modifying the p_interp
source
or the MET code to do that rotation.
Or you could write a stand-alone script to process the output of
p_interp
and rotate the winds before passing them to MET.
Sorry I don't have an easier answer for you.
Thanks,
John
------------------------------------------------
More information about the Met_help
mailing list