[Met_help] [rt.rap.ucar.edu #42772] History for WaveletStat error
RAL HelpDesk {for John Halley Gotway}
met_help at ucar.edu
Wed Dec 15 13:36:53 MST 2010
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
I recently (today) switched over to METv3.0 from METv2.0 and updated the WaveletStatConfig
file to match the default file in the METv3.0 folder, but I get an error I've never seen before.
This is what I get:
Forecast File: forecastw.grb
Observation File: observationw.nc
Configuration File: WaveletStatConfig_process
wavelet_stat(5965): unaligned access to 0x60000fffff9c501c, ip=0x4000000000168a81
wavelet_stat(5965): unaligned access to 0x60000fffff9c501c, ip=0x4000000000168a81
wavelet_stat(5965): unaligned access to 0x60000fffff9c4f4c, ip=0x4000000000168a81
wavelet_stat(5965): unaligned access to 0x60000fffff9c457c, ip=0x4000000000168a81
MetNcFile::get_times(const NcVar *) -> can't get init and/or valid times for variable LAT
Any help would be much appreciated. I attached my configuration file and the obs and
forecast files.
Thanks,
Derek Stratman
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: WaveletStat error
From: John Halley Gotway
Time: Tue Dec 14 14:25:23 2010
Derek,
Thanks for sending that sample data. I was able to reproduce the
issue you're having. Once I made a handful of changes to your NetCDF
file, I was able to run it through the METv3.0 version of
Wavelet-Stat.
Here are the required changes:
(1) Rename the variable "LAT" to "lat" and rename the variable "LON"
to "lon". Unfortunately, the checking of the variable names in MET is
case sensitive right now. MET expects those variables to be
lowercase. Once you rename them, you'll get past that error.
(2) Next, unfortunately, the grid specs defined in the NetCDF header
do not EXACTLY match the grid definition in the GRIB file:
NetCDF grid definition (ncdump -h):
:ny = "256 grid_points" ;
:nx = "256 grid_points" ;
:r_km = "6367.470000 km" ;
:d_km = "4.762500 km" ;
:lcen_deg = "-105.000000 degrees_east" ;
:p1_deg = "39.185001 degrees_north" ;
:p0_deg = "30.978001 degrees_north" ;
:l0_deg = "-105.000000 degrees_east" ;
:Projection = "Polar Stereographic" ;
GRIB grid definition (wgrib -V):
polar stereo: Lat1 30.978000 Long1 -105.000000 Orient
-105.000000
north pole (256 x 256) Dx 4763 Dy 4763 scan 64 mode 8
In order to figure out what MET was looking for, I actually took your
GRIB file and ran it through the pcp_combine (METv2.0) tool:
METv2.0_patch/bin/pcp_combine -gc 212 -add
caps_cn_2010050115f000_256x256.grb 0
caps_cn_2010050115f000_256x256.grb 0 caps.nc
Here's the resulting grid definition (ncdump -h caps.nc):
:Projection = "Polar Stereographic" ;
:p1_deg = "60.000000 degrees_north" ;
:p0_deg = "30.978000 degrees_north" ;
:l0_deg = "-105.000000 degrees_east" ;
:lcen_deg = "-105.000000 degrees_east" ;
:d_km = "4.763000 km" ;
:r_km = "6367.470000 km" ;
:nx = "256 grid_points" ;
:ny = "256 grid_points" ;
So I went into your NetCDF file and made the following changes:
- Change "39.185001 degrees_north" to "39.185 degrees_north"
- Change "30.978001 degrees_north" to "30.978 degrees_north"
- Change "4.762500 km" to "4.763 km"
- Change "39.185001 degrees_north" to "60.0 degrees_north"
The modified file is attached. Once I made those changes, the METv3.0
version of Wavelet-Stat was able to read it.
Two last comments...
- I don't know why your grid definition differs between your NetCDF
file and your GRIB file, but you should look into that.
- Please keep in mind that when running Wavelet-Stat you actually do
NOT need to first put your data onto a dyadic grid. Instead, you can
pass it data on any grid and then use the Wavelet-Stat
configuration options to define the subset of grid-points over which
you'd like to run the tools using the following parameters:
grid_decomp_flag, tile_xll, tile_yll, and tile_dim
Hope that helps.
John Halley Gotway
On 12/14/2010 12:32 PM, RAL HelpDesk {for Derek.Stratman} wrote:
>
> Tue Dec 14 12:32:24 2010: Request 42772 was acted upon.
> Transaction: Ticket created by Derek.Stratman at noaa.gov
> Queue: met_help
> Subject: WaveletStat error
> Owner: Nobody
> Requestors: Derek.Stratman at noaa.gov
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=42772 >
>
>
> I recently (today) switched over to METv3.0 from METv2.0 and updated
the WaveletStatConfig
> file to match the default file in the METv3.0 folder, but I get an
error I've never seen before.
> This is what I get:
>
> Forecast File: forecastw.grb
> Observation File: observationw.nc
> Configuration File: WaveletStatConfig_process
> wavelet_stat(5965): unaligned access to 0x60000fffff9c501c,
ip=0x4000000000168a81
> wavelet_stat(5965): unaligned access to 0x60000fffff9c501c,
ip=0x4000000000168a81
> wavelet_stat(5965): unaligned access to 0x60000fffff9c4f4c,
ip=0x4000000000168a81
> wavelet_stat(5965): unaligned access to 0x60000fffff9c457c,
ip=0x4000000000168a81
>
>
> MetNcFile::get_times(const NcVar *) -> can't get init and/or valid
times for variable LAT
>
> Any help would be much appreciated. I attached my configuration
file and the obs and
> forecast files.
>
> Thanks,
> Derek Stratman
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #42772] WaveletStat error
From: Derek.Stratman
Time: Wed Dec 15 12:06:59 2010
Hi John,
Thanks for your help. I didn't spot the problem when I ran gridstat
with METv2.0, but after
reviewing the output in more depth, it's obvious the problems you
found are the cause. I'll
let you know if I find anything else wrong once these changes are
made, but for now, I'm
good to go.
Thanks again,
Derek Stratman
----- Original Message -----
From: RAL HelpDesk {for John Halley Gotway} <met_help at ucar.edu>
Date: Tuesday, December 14, 2010 3:25 pm
Subject: Re: [rt.rap.ucar.edu #42772] WaveletStat error
> Derek,
>
> Thanks for sending that sample data. I was able to reproduce the
> issue you're having. Once I made a handful of changes to your
> NetCDF file, I was able to run it through the METv3.0 version of
> Wavelet-Stat.
>
> Here are the required changes:
> (1) Rename the variable "LAT" to "lat" and rename the variable
> "LON" to "lon". Unfortunately, the checking of the variable names
> in MET is case sensitive right now. MET expects those variables to
be
> lowercase. Once you rename them, you'll get past that error.
>
> (2) Next, unfortunately, the grid specs defined in the NetCDF
> header do not EXACTLY match the grid definition in the GRIB file:
> NetCDF grid definition (ncdump -h):
> :ny = "256 grid_points" ;
> :nx = "256 grid_points" ;
> :r_km = "6367.470000 km" ;
> :d_km = "4.762500 km" ;
> :lcen_deg = "-105.000000 degrees_east" ;
> :p1_deg = "39.185001 degrees_north" ;
> :p0_deg = "30.978001 degrees_north" ;
> :l0_deg = "-105.000000 degrees_east" ;
> :Projection = "Polar Stereographic" ;
> GRIB grid definition (wgrib -V):
> polar stereo: Lat1 30.978000 Long1 -105.000000 Orient -
> 105.000000 north pole (256 x 256) Dx 4763 Dy 4763 scan 64
> mode 8
>
> In order to figure out what MET was looking for, I actually took
> your GRIB file and ran it through the pcp_combine (METv2.0) tool:
> METv2.0_patch/bin/pcp_combine -gc 212 -add
> caps_cn_2010050115f000_256x256.grb 0
> caps_cn_2010050115f000_256x256.grb 0 caps.nc
>
> Here's the resulting grid definition (ncdump -h caps.nc):
> :Projection = "Polar Stereographic" ;
> :p1_deg = "60.000000 degrees_north" ;
> :p0_deg = "30.978000 degrees_north" ;
> :l0_deg = "-105.000000 degrees_east" ;
> :lcen_deg = "-105.000000 degrees_east" ;
> :d_km = "4.763000 km" ;
> :r_km = "6367.470000 km" ;
> :nx = "256 grid_points" ;
> :ny = "256 grid_points" ;
>
> So I went into your NetCDF file and made the following changes:
> - Change "39.185001 degrees_north" to "39.185 degrees_north"
> - Change "30.978001 degrees_north" to "30.978 degrees_north"
> - Change "4.762500 km" to "4.763 km"
> - Change "39.185001 degrees_north" to "60.0 degrees_north"
>
> The modified file is attached. Once I made those changes, the
> METv3.0 version of Wavelet-Stat was able to read it.
>
> Two last comments...
> - I don't know why your grid definition differs between your NetCDF
> file and your GRIB file, but you should look into that.
> - Please keep in mind that when running Wavelet-Stat you actually
> do NOT need to first put your data onto a dyadic grid. Instead,
> you can pass it data on any grid and then use the Wavelet-Stat
> configuration options to define the subset of grid-points over
> which you'd like to run the tools using the following parameters:
> grid_decomp_flag, tile_xll, tile_yll, and tile_dim
>
> Hope that helps.
>
> John Halley Gotway
>
> On 12/14/2010 12:32 PM, RAL HelpDesk {for Derek.Stratman} wrote:
> >
> > Tue Dec 14 12:32:24 2010: Request 42772 was acted upon.
> > Transaction: Ticket created by Derek.Stratman at noaa.gov
> > Queue: met_help
> > Subject: WaveletStat error
> > Owner: Nobody
> > Requestors: Derek.Stratman at noaa.gov
> > Status: new
> > Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=42772 >
> >
> >
> > I recently (today) switched over to METv3.0 from METv2.0 and
> updated the WaveletStatConfig
> > file to match the default file in the METv3.0 folder, but I get
> an error I've never seen before.
> > This is what I get:
> >
> > Forecast File: forecastw.grb
> > Observation File: observationw.nc
> > Configuration File: WaveletStatConfig_process
> > wavelet_stat(5965): unaligned access to 0x60000fffff9c501c,
> ip=0x4000000000168a81> wavelet_stat(5965): unaligned access to
> 0x60000fffff9c501c, ip=0x4000000000168a81
> > wavelet_stat(5965): unaligned access to 0x60000fffff9c4f4c,
> ip=0x4000000000168a81> wavelet_stat(5965): unaligned access to
> 0x60000fffff9c457c, ip=0x4000000000168a81
> >
> >
> > MetNcFile::get_times(const NcVar *) -> can't get init and/or
> valid times for variable LAT
> >
> > Any help would be much appreciated. I attached my configuration
> file and the obs and
> > forecast files.
> >
> > Thanks,
> > Derek Stratman
> >
> >
>
>
------------------------------------------------
More information about the Met_help
mailing list