[Met_help] MET/pcp_combine help

John Halley Gotway johnhg at ucar.edu
Wed Feb 3 12:07:58 MST 2010 

Jeff,

OK.  I'll hold off on working on any changes to pcp_combine to handle off-hour accumulations in this way.  If this does turn out to be an issue down the road, please let us know, and we could probably
put something together for you.

Thanks,
John

Jeffrey Duda wrote:
> John,
> Thanks for the elaborate explanation.  It helps me better understand how
> pcp_combine works.
> 
> This is not really a pressing issue for me.  I don't have observational data
> that is off-hour.  The data I was thinking of was my own WRF simulations
> where I have output files every 15 minutes.  I'm using pcp_combine to get a
> 1-hr precip data set to compare to Stage IV data to run MODE for my model
> runs.  It turns out that the MCS in my simulation could be best described as
> initiating between 20:15 and 21:15 UTC on a particular day, so I was going
> to see if I could use that in my comparsion.  It looks like I cannot right
> now.  I can use 2000 UTC to 2100 UTC just fine.  Thanks, though.
> 
> Jeff
> 
> On Wed, Feb 3, 2010 at 12:24 PM, John Halley Gotway <johnhg at ucar.edu> wrote:
> 
>> Jeff,
>>
>> Very good question.  And I think the answer is that pcp_combine can't
>> currently do that.  The "subtract" option for pcp_combine only allows you to
>> specify the HH of the accumulation interval, not
>> HHMMSS.  So you'd only be able to perform the subtraction at the top of
>> each hour.
>>
>> Enhancing pcp_combine to allow you to specify HHHMMSS would be pretty
>> straight-forward, but there would be implications for other MET tools.
>>  Right now, pcp_combine creates a NetCDF file where the
>> variables for precip are named as "APCP_HH" where HH indicates the
>> accumulation interval.  If we allow off-hour forecasts to be used, we'd need
>> to generalize the variable names from "APCP_HH" to
>> "APCP_HHMMSS".  So we'd need to enhance the other MET tools to read that
>> data.
>>
>> Is this a pressing need for you?  Do you need to combine off-hour data in
>> this way for your verification?  I'm also curious about your observations.
>>  Do you really have observed precip that's
>> accumulated to 15 minutes after the hour?
>>
>> If this is an issue that's standing in your way of using MET, we could
>> probably figure out a fix.  We could enhance the code to allow you to
>> specify HHMMSS for pcp_combine.  If you do need to pursue
>> this, please send me some sample data that includes some off-hour
>> accumulations and a description of the pcp_combine function you'd like to
>> perform.
>>
>> Thanks,
>> John
>>
>>
>> Jeffrey Duda wrote:
>>> John,
>>> I have one more question for you.  If I wanted to use pcp_combine to get
>>> hourly precipitation from run total accumulated precip from WRFPRS files
>>> that are off-hour (like WRFPRS_d02.008_15), how would I indicate that
>> when
>>> typing in the subtract args in the pcp_combine -subtract command?
>>>
>>> Jeff Duda
>>>
>>> On Sun, Jan 3, 2010 at 8:46 PM, John Halley Gotway <johnhg at rap.ucar.edu
>>> wrote:
>>>
>>>> Jeff,
>>>>
>>>> Sorry for the delay in getting back to you.
>>>>
>>>> Actually, the output of the PCP-Combine tool is a NetCDF file.  We use
>>>> NetCDF as our internal file format.  The other MET tools, like
>> Grid-Stat,
>>>> Point-Stat, MODE, and Wavelet-Stat are all able to read GRIB files or
>> the
>>>> NetCDF output of the PCP-Combine tool.
>>>>
>>>> For the next release of MET, we're working on generalizing our support
>> for
>>>> NetCDF so that we're able to read more flavors of NetCDF and not just
>> the
>>>> output of the PCP-Combine tool.
>>>>
>>>> Since the output is NetCDF, I'd suggest using the ncdump or ncview tools
>>>> to display/interrogate the data.  And let me refer to the MET User's
>> Guide
>>>> (
>>>>
>> http://www.dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v2.0_rev2.pdf
>>>> )
>>>> for more details.
>>>>
>>>> Hope that helps.
>>>>
>>>> John
>>>>
>>>>> John,
>>>>> I was able to successfully run copygb with your help and now am looking
>>>> to
>>>>> run pcp_combine, subtract, to get a one hour precip field from my WRF
>>>>> output
>>>>> files.  I have isolated the two files (and actually uploaded them to
>> the
>>>>> server described in previous emails), and used the following command to
>>>>> run
>>>>> pcp_combine assuming the following:
>>>>>
>>>>>   "accum1" indicates the accumulation interval to be used from in_file1
>>>> in
>>>>> HH format (required).   -I set this as 12
>>>>>   "accum2" indicates the accumulation interval to be used from in_file2
>>>> in
>>>>> HH format (required).   -I set this as 11
>>>>>
>>>>> I got this:
>>>>>
>>>>> /bin/pcp_combine -subtract WRFPRS_d02.012_00 12 WRFPRS_d02.011_00 11
>>>>> difference.out
>>>>> Reading input file: WRFPRS_d02.012_00
>>>>> Reading input file: WRFPRS_d02.011_00
>>>>> Performing subtraction command.
>>>>> Writing output file: difference.out
>>>>>
>>>>> Figuring that pcp_combine took the difference between two GRIB files,
>> the
>>>>> output should also be grip, no?  So I tried to convert it to view using
>>>>> Grads, but got an error when I ran grib2ctl.pl:
>>>>> "/grib2ctl.pl difference.out > difference_ctl.ctl
>>>>> found unidentified data
>>>>> Big problem:
>>>>>   either difference.out is missing or not a grib file
>>>>>   or wgrib is not on your path
>>>>>
>>>>> Please help.  Thank you very much for your help so far.
>>>>>
>>>>> Jeff Duda
>>>>>
>>>>> On Fri, Dec 18, 2009 at 3:45 PM, John Halley Gotway
>>>>> <johnhg at rap.ucar.edu>wrote:
>>>>>
>>>>>> Jeff,
>>>>>>
>>>>>> Well I'd first check to see what's actually in that GRIB file using
>> the
>>>>>> "wgrib" command.  For example, try this:
>>>>>> wgrib WRFPRS_d02.012_00 | grep APCP
>>>>>>
>>>>>> That'll dump out the contents of that GRIB file, looking only for the
>>>>>> APCP
>>>>>> records (accumulated precip).  Look to see if that file really does
>>>>>> contain a record for 0-12 hour accumulated precip.
>>>>>>
>>>>>> If you have more questions about this, it'd probably be easier for you
>>>>>> to
>>>>>> just send me some sample data to take a look at.  If you'd like to do
>>>>>> that, you could post it to our anonymous ftp site in the following
>> way:
>>>>>> cd "the directory that contains the files you want to send"
>>>>>> ftp ftp.rap.ucar.edu
>>>>>> username = anonymous
>>>>>> password = "your email address"
>>>>>> cd incoming/irap/met_help
>>>>>> mkdir duda_data
>>>>>> cd duda_data
>>>>>> put WRFPRS_d02.012_00
>>>>>> put WRFPRS_d02.011_00
>>>>>> bye
>>>>>>
>>>>>> Hope that helps.
>>>>>>
>>>>>> John
>>>>>>
>>>>>>> John,
>>>>>>> I am Bill Gallus' student and I am trying to use MODE to verify a WRF
>>>>>>> model
>>>>>>> run using accumulated precip.  To do this, I need to match Stage_IV
>>>>>> data
>>>>>>> to
>>>>>>> my model grid.  To do that, though, I need to get a difference in
>>>>>> precip
>>>>>>> file to compare with Stage_IV data (which has 1 hourly accumulated
>>>>>>> precip).
>>>>>>> My WRF files use run total accumulated precip.  I have run WPP on the
>>>>>>> NetCDF
>>>>>>> output of the WRF files and now am trying to use pcp_combine
>> -subtract
>>>>>> to
>>>>>>> basically subtract the total accumulated precip from one file from
>> one
>>>>>> an
>>>>>>> hour before it (my NetCDF output interval is 15 minutes).  When I
>>>>>> tried
>>>>>> to
>>>>>>> run pcp_combine, though, I got an error.  Can you help me with this,
>>>>>>> please?  Thank you.
>>>>>>>
>>>>>>>
>>>>>>> Here's a screen grab of the command and error:
>>>>>>>
>>>>>>> -bash-3.2$ /usr/local/METv2.0/bin/pcp_combine -subtract
>>>>>> WRFPRS_d02.012_00
>>>>>>> 12
>>>>>>> WRFPRS_d02.011_00 11 20080719_18-17_precipdiff
>>>>>>> Reading input file: WRFPRS_d02.012_00
>>>>>>>
>>>>>>>
>>>>>>> ERROR: get_field() -> can't find grib code 61 with accumulation of 12
>>>>>>> hours
>>>>>>> in grib file: WRFPRS_d02.012_00
>>>>>>>
>>>>>>> Jeff Duda
>>>>>>> --
>>>>>>> Jeff Duda
>>>>>>> Iowa State University
>>>>>>> Meteorology Graduate Student
>>>>>>> 3134 Agronomy Hall
>>>>>>> www.meteor.iastate.edu/~jdduda<http://www.meteor.iastate.edu/%7Ejdduda>
>> <http://www.meteor.iastate.edu/%7Ejdduda>
>>>>>> <http://www.meteor.iastate.edu/%7Ejdduda>
>>>>>>
>>>>>>
>>>>> --
>>>>> Jeff Duda
>>>>> Iowa State University
>>>>> Meteorology Graduate Student
>>>>> 3134 Agronomy Hall
>>>>> www.meteor.iastate.edu/~jdduda<http://www.meteor.iastate.edu/%7Ejdduda><
>> http://www.meteor.iastate.edu/%7Ejdduda>
>>>>> _______________________________________________
>>>>> Met_help mailing list
>>>>> Met_help at mailman.ucar.edu
>>>>> http://mailman.ucar.edu/mailman/listinfo/met_help
>>>>>
>>>>
>>>
> 
> 
>

More information about the Met_help mailing list