[Met_help] [rt.rap.ucar.edu #65939] History for Help
John Halley Gotway via RT
met_help at ucar.edu
Mon Jun 2 14:15:43 MDT 2014
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Good afternoon met_help,
I am trying to use MET to verify WRF runs against a locally generated NEXRAD precipitation estimate. I use matlab so the local precipitation data is not in GRIB. I did however figure out how to create a netcdf from the data that matched the output from pcp_combine. The problem is that my data is lat/lon based so I set up the netcdf with the appropriate lat/lon attributes to match my data. When I ran grid_stat and mode on the data I got this error: read_netcdf_grid_v3() -> Projection type Lat/Lon not currently supported.
I used lat/lon because the tutorial and handbook say that lat/lon is supported.
Should I try to reference my data to another projection? Is there a better way to get my data formatted as an input for grid_stat and mode?
I can provide files etc. if your require them.
Thank you for your time,
John D McMillen
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Help
From: John Halley Gotway
Time: Fri Mar 21 16:19:15 2014
John,
I see that you're having trouble getting gridded data formatted in a
way for the MET tools to read them.
Yes, the lat/lon projection is supported. I suspect there's a minor
formatting issue causing the problem. Please check the NetCDF file
you created to see if the "Projection" attribute is set exactly
to "LatLon". If, for example, it was set to "lat/lon", you'd get the
error you're seeing.
If that fixes the problem, great. If not, please send me a sample
NetCDF file and let me know what version of MET you're running. You
can post data to our anonymous ftp site following these
instructions:
http://www.dtcenter.org/met/users/support/met_help.php#ftp
Hope that helps.
Thanks,
John
On 03/21/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>
> Fri Mar 21 15:21:32 2014: Request 65939 was acted upon.
> Transaction: Ticket created by john.mcmillen at utah.edu
> Queue: met_help
> Subject: Help
> Owner: Nobody
> Requestors: john.mcmillen at utah.edu
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>
>
> Good afternoon met_help,
>
> I am trying to use MET to verify WRF runs against a locally
generated NEXRAD precipitation estimate. I use matlab so the local
precipitation data is not in GRIB. I did however figure out how to
create a netcdf from the data that matched the output from
pcp_combine. The problem is that my data is lat/lon based so I set up
the netcdf with the appropriate lat/lon attributes to match my data.
When I ran grid_stat and mode on the data I got this error:
read_netcdf_grid_v3() -> Projection type Lat/Lon not currently
supported.
>
> I used lat/lon because the tutorial and handbook say that lat/lon is
supported.
>
> Should I try to reference my data to another projection? Is there a
better way to get my data formatted as an input for grid_stat and
mode?
>
> I can provide files etc. if your require them.
>
> Thank you for your time,
>
> John D McMillen
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Help
From: JOHN DANIEL MCMILLEN
Time: Fri Mar 21 16:28:43 2014
Wow,
Great response time!
You are correct, I added a slash and that caused the error.
Thanks,
John
On Mar 21, 2014, at 4:19 PM, John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> John,
>
> I see that you're having trouble getting gridded data formatted in a
way for the MET tools to read them.
>
> Yes, the lat/lon projection is supported. I suspect there's a minor
formatting issue causing the problem. Please check the NetCDF file
you created to see if the "Projection" attribute is set exactly
> to "LatLon". If, for example, it was set to "lat/lon", you'd get
the error you're seeing.
>
> If that fixes the problem, great. If not, please send me a sample
NetCDF file and let me know what version of MET you're running. You
can post data to our anonymous ftp site following these
> instructions:
> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>
> Hope that helps.
>
> Thanks,
> John
>
> On 03/21/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>
>> Fri Mar 21 15:21:32 2014: Request 65939 was acted upon.
>> Transaction: Ticket created by john.mcmillen at utah.edu
>> Queue: met_help
>> Subject: Help
>> Owner: Nobody
>> Requestors: john.mcmillen at utah.edu
>> Status: new
>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>
>>
>> Good afternoon met_help,
>>
>> I am trying to use MET to verify WRF runs against a locally
generated NEXRAD precipitation estimate. I use matlab so the local
precipitation data is not in GRIB. I did however figure out how to
create a netcdf from the data that matched the output from
pcp_combine. The problem is that my data is lat/lon based so I set up
the netcdf with the appropriate lat/lon attributes to match my data.
When I ran grid_stat and mode on the data I got this error:
read_netcdf_grid_v3() -> Projection type Lat/Lon not currently
supported.
>>
>> I used lat/lon because the tutorial and handbook say that lat/lon
is supported.
>>
>> Should I try to reference my data to another projection? Is there
a better way to get my data formatted as an input for grid_stat and
mode?
>>
>> I can provide files etc. if your require them.
>>
>> Thank you for your time,
>>
>> John D McMillen
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Help
From: John Halley Gotway
Time: Mon Mar 24 09:58:35 2014
Great, glad that did the trick. Thanks for letting me know.
Just let us know if any more issues or questions come up in your use
of MET.
Thanks,
John
On 03/21/2014 04:28 PM, JOHN DANIEL MCMILLEN via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>
> Wow,
>
> Great response time!
>
> You are correct, I added a slash and that caused the error.
>
> Thanks,
>
> John
>
> On Mar 21, 2014, at 4:19 PM, John Halley Gotway via RT
<met_help at ucar.edu>
> wrote:
>
>> John,
>>
>> I see that you're having trouble getting gridded data formatted in
a way for the MET tools to read them.
>>
>> Yes, the lat/lon projection is supported. I suspect there's a
minor formatting issue causing the problem. Please check the NetCDF
file you created to see if the "Projection" attribute is set exactly
>> to "LatLon". If, for example, it was set to "lat/lon", you'd get
the error you're seeing.
>>
>> If that fixes the problem, great. If not, please send me a sample
NetCDF file and let me know what version of MET you're running. You
can post data to our anonymous ftp site following these
>> instructions:
>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>
>> Hope that helps.
>>
>> Thanks,
>> John
>>
>> On 03/21/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>>
>>> Fri Mar 21 15:21:32 2014: Request 65939 was acted upon.
>>> Transaction: Ticket created by john.mcmillen at utah.edu
>>> Queue: met_help
>>> Subject: Help
>>> Owner: Nobody
>>> Requestors: john.mcmillen at utah.edu
>>> Status: new
>>> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>
>>>
>>> Good afternoon met_help,
>>>
>>> I am trying to use MET to verify WRF runs against a locally
generated NEXRAD precipitation estimate. I use matlab so the local
precipitation data is not in GRIB. I did however figure out how to
create a netcdf from the data that matched the output from
pcp_combine. The problem is that my data is lat/lon based so I set up
the netcdf with the appropriate lat/lon attributes to match my data.
When I ran grid_stat and mode on the data I got this error:
read_netcdf_grid_v3() -> Projection type Lat/Lon not currently
supported.
>>>
>>> I used lat/lon because the tutorial and handbook say that lat/lon
is supported.
>>>
>>> Should I try to reference my data to another projection? Is there
a better way to get my data formatted as an input for grid_stat and
mode?
>>>
>>> I can provide files etc. if your require them.
>>>
>>> Thank you for your time,
>>>
>>> John D McMillen
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: JOHN DANIEL MCMILLEN
Time: Wed Mar 26 15:21:25 2014
Hello MET Help,
I am using mode to analyze NEXRAD accumulations and WRF accumulations
that I am formatting as pcp_combine netcdf files. I am using the
NEXRAD latlon grid as the common grid and have filled out the netcdf
attributes as such. The tool is working and the analysis seems to be
correct. The issue I have is the data is not plotted correctly with
respect to the maps in the .ps file output by mode. It appears that
the data is being plotted ~20km east of the plotted map.
When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
Thanks for your time,
John D McMillen
On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> According to our records, your request has been resolved. If you
have any
> further questions or concerns, please respond to this message.
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: John Halley Gotway
Time: Wed Mar 26 17:15:02 2014
John,
Yes, it'd help to look at some data files. You could also try running
them through the "plot_data_plane" utility, which would take MODE out
of the equation. It's likely an issue in how the grid is
specified. Please send a sample NetCDF file, the PostScript output of
MODE, and a plot from Matlab showing where the data "should" be.
If the files are big, you could post them to our anonymous ftp site
following these instructions:
http://www.dtcenter.org/met/users/support/met_help.php#ftp
Thanks,
John
On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>
> Hello MET Help,
>
> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>
> When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
>
> Thanks for your time,
>
> John D McMillen
>
>
> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
> wrote:
>
>> According to our records, your request has been resolved. If you
have any
>> further questions or concerns, please respond to this message.
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: JOHN DANIEL MCMILLEN
Time: Wed Mar 26 17:42:07 2014
Thanks John.
I will send them tomorrow am.
John
Sent from my iPhone
> On Mar 26, 2014, at 5:15 PM, "John Halley Gotway via RT"
<met_help at ucar.edu> wrote:
>
> John,
>
> Yes, it'd help to look at some data files. You could also try
running them through the "plot_data_plane" utility, which would take
MODE out of the equation. It's likely an issue in how the grid is
> specified. Please send a sample NetCDF file, the PostScript output
of MODE, and a plot from Matlab showing where the data "should" be.
>
> If the files are big, you could post them to our anonymous ftp site
following these instructions:
> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>
> Thanks,
> John
>
>> On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>
>> Hello MET Help,
>>
>> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>>
>> When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
>>
>> Thanks for your time,
>>
>> John D McMillen
>>
>>
>> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>> wrote:
>>
>>> According to our records, your request has been resolved. If you
have any
>>> further questions or concerns, please respond to this message.
>
------------------------------------------------
Subject: Help
From: JOHN DANIEL MCMILLEN
Time: Thu Mar 27 09:10:21 2014
John,
Here are the files. The two .nc files are the inout files I use in
MODE. WRF is the forecast, NEXRAD is the observation. I based the
format for each on the attributes of pcp_combine NetCDF files from the
tutorial. The .ps file is the output of MODE. I believe the analysis
is accurate because the placement of the forecast and observed data is
accurate relative to one another as they are on the same grid.
Comparing the images from the .ps to the .png I created from the input
files shows the same relative placement of observation to forecast,
but shows that both fields should be over the southern portion of the
Great Salt Lake, not shifted eastward. In the .png the contour is the
same as conv_thresh in my config file. The blue contour is the
forecast and the black contour is the observation.
[cid:418d776b-a30d-422c-9d1b-daba71a8b2c0 at umail.utah.edu]
Let me know if you have any other questions.
Thanks for your help,
John
On Mar 26, 2014, at 5:15 PM, John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> John,
>
> Yes, it'd help to look at some data files. You could also try
running them through the "plot_data_plane" utility, which would take
MODE out of the equation. It's likely an issue in how the grid is
> specified. Please send a sample NetCDF file, the PostScript output
of MODE, and a plot from Matlab showing where the data "should" be.
>
> If the files are big, you could post them to our anonymous ftp site
following these instructions:
> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>
> Thanks,
> John
>
> On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>
>> Hello MET Help,
>>
>> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>>
>> When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
>>
>> Thanks for your time,
>>
>> John D McMillen
>>
>>
>> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>> wrote:
>>
>>> According to our records, your request has been resolved. If you
have any
>>> further questions or concerns, please respond to this message.
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: John Halley Gotway
Time: Thu Mar 27 10:23:38 2014
John,
Using ncdump, I see the following in the NetCDF files you sent:
ncdump -h NEXRAD20101027.nc | egrep "lat_ll|lon_ll"
:lat_ll = 40.1007299270073 ;
:lon_ll = -113.69625 ;
ncdump -h WRF20101027.nc | egrep "lat_ll|lon_ll"
:lat_ll = "40.1011" ;
:lon_ll = "-113.4123" ;
As you can see, that latitudes are very close, but the longitudes
differ by .28395 degrees. That explains the 20km or so offset you're
seeing. The real mystery now is why MODE allowed you to compare
the files in the first place!
When I tried to run these two files through MODE from METv4.1, I got
the following error:
ERROR : process_fcst_obs_files() -> The forecast and observation
grids do not match: Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101
lon_ll: 113.412 delta_lat: 0.009 delta_lon: 0.009 !=
Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101 lon_ll: 113.696
delta_lat: 0.009 delta_lon: 0.009
Why didn't you? What version of MET are you running? Have you made
any changes to the source code yourself?
FYI, in the process of debugging your issue, I used the
plot_data_plane utility to make plots of the NEXRAD and WRF data
directly. But since it's such a small area, I decided to turn on the
plotting
of the USA counties. Here's how:
(1) Edit the file METv4.1/data/config/ConfigMapData
In the "source" section, add a 4th entry for the file
"usa_county_data":
source = [
{
file_name = "MET_BASE/data/map/legacy/world_data_minus_usa";
},
{
file_name =
"MET_BASE/data/map/legacy/country_data_minus_usa";
},
{
file_name = "MET_BASE/data/map/legacy/usa_state_data";
},
{
file_name = "MET_BASE/data/map/legacy/usa_county_data";
}
];
Now, all of the MET tools that make plots (like MODE, plot_data_plane,
and plot_point_obs) will include the USA county data in their plots.
(2) Then run plot_data_plane like this:
METv4.1/bin/plot_data_plane NEXRAD20101027.nc NEXRAD20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
METv4.1/bin/plot_data_plane WRF20101027.nc WRF20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
That'll plot the data directly, each on it's own grid. I've specified
the plotting range to make them consistent with the plots from MODE
you sent. You'll see that the resulting grids really are
different. That shift due to the longitude difference is pretty
obvious.
Hope that helps.
Thanks,
John
On 03/27/2014 09:10 AM, JOHN DANIEL MCMILLEN via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>
> John,
>
> Here are the files. The two .nc files are the inout files I use in
MODE. WRF is the forecast, NEXRAD is the observation. I based the
format for each on the attributes of pcp_combine NetCDF files from the
tutorial. The .ps file is the output of MODE. I believe the analysis
is accurate because the placement of the forecast and observed data is
accurate relative to one another as they are on the same grid.
Comparing the images from the .ps to the .png I created from the input
files shows the same relative placement of observation to forecast,
but shows that both fields should be over the southern portion of the
Great Salt Lake, not shifted eastward. In the .png the contour is the
same as conv_thresh in my config file. The blue contour is the
forecast and the black contour is the observation.
>
>
> [cid:418d776b-a30d-422c-9d1b-daba71a8b2c0 at umail.utah.edu]
>
>
> Let me know if you have any other questions.
>
> Thanks for your help,
> John
>
> On Mar 26, 2014, at 5:15 PM, John Halley Gotway via RT
<met_help at ucar.edu>
> wrote:
>
>> John,
>>
>> Yes, it'd help to look at some data files. You could also try
running them through the "plot_data_plane" utility, which would take
MODE out of the equation. It's likely an issue in how the grid is
>> specified. Please send a sample NetCDF file, the PostScript output
of MODE, and a plot from Matlab showing where the data "should" be.
>>
>> If the files are big, you could post them to our anonymous ftp site
following these instructions:
>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>
>> Thanks,
>> John
>>
>> On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>
>>> Hello MET Help,
>>>
>>> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>>>
>>> When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
>>>
>>> Thanks for your time,
>>>
>>> John D McMillen
>>>
>>>
>>> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>>> wrote:
>>>
>>>> According to our records, your request has been resolved. If you
have any
>>>> further questions or concerns, please respond to this message.
>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: John Halley Gotway
Time: Thu Mar 27 10:24:21 2014
John,
Using ncdump, I see the following in the NetCDF files you sent:
ncdump -h NEXRAD20101027.nc | egrep "lat_ll|lon_ll"
:lat_ll = 40.1007299270073 ;
:lon_ll = -113.69625 ;
ncdump -h WRF20101027.nc | egrep "lat_ll|lon_ll"
:lat_ll = "40.1011" ;
:lon_ll = "-113.4123" ;
As you can see, that latitudes are very close, but the longitudes
differ by .28395 degrees. That explains the 20km or so offset you're
seeing. The real mystery now is why MODE allowed you to compare
the files in the first place!
When I tried to run these two files through MODE from METv4.1, I got
the following error:
ERROR : process_fcst_obs_files() -> The forecast and observation
grids do not match: Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101
lon_ll: 113.412 delta_lat: 0.009 delta_lon: 0.009 !=
Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101 lon_ll: 113.696
delta_lat: 0.009 delta_lon: 0.009
Why didn't you? What version of MET are you running? Have you made
any changes to the source code yourself?
FYI, in the process of debugging your issue, I used the
plot_data_plane utility to make plots of the NEXRAD and WRF data
directly. But since it's such a small area, I decided to turn on the
plotting
of the USA counties. Here's how:
(1) Edit the file METv4.1/data/config/ConfigMapData
In the "source" section, add a 4th entry for the file
"usa_county_data":
source = [
{
file_name = "MET_BASE/data/map/legacy/world_data_minus_usa";
},
{
file_name =
"MET_BASE/data/map/legacy/country_data_minus_usa";
},
{
file_name = "MET_BASE/data/map/legacy/usa_state_data";
},
{
file_name = "MET_BASE/data/map/legacy/usa_county_data";
}
];
Now, all of the MET tools that make plots (like MODE, plot_data_plane,
and plot_poitn
(2) Then run plot_data_plane like this:
METv4.1/bin/plot_data_plane NEXRAD20101027.nc NEXRAD20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
METv4.1/bin/plot_data_plane WRF20101027.nc WRF20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
That'll plot the data directly, each on it's own grid. And you'll see
that the resulting grids really are shifted.
Hope that helps.
Thanks,
John
On 03/27/2014 09:10 AM, JOHN DANIEL MCMILLEN via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>
> John,
>
> Here are the files. The two .nc files are the inout files I use in
MODE. WRF is the forecast, NEXRAD is the observation. I based the
format for each on the attributes of pcp_combine NetCDF files from the
tutorial. The .ps file is the output of MODE. I believe the analysis
is accurate because the placement of the forecast and observed data is
accurate relative to one another as they are on the same grid.
Comparing the images from the .ps to the .png I created from the input
files shows the same relative placement of observation to forecast,
but shows that both fields should be over the southern portion of the
Great Salt Lake, not shifted eastward. In the .png the contour is the
same as conv_thresh in my config file. The blue contour is the
forecast and the black contour is the observation.
>
>
> [cid:418d776b-a30d-422c-9d1b-daba71a8b2c0 at umail.utah.edu]
>
>
> Let me know if you have any other questions.
>
> Thanks for your help,
> John
>
> On Mar 26, 2014, at 5:15 PM, John Halley Gotway via RT
<met_help at ucar.edu>
> wrote:
>
>> John,
>>
>> Yes, it'd help to look at some data files. You could also try
running them through the "plot_data_plane" utility, which would take
MODE out of the equation. It's likely an issue in how the grid is
>> specified. Please send a sample NetCDF file, the PostScript output
of MODE, and a plot from Matlab showing where the data "should" be.
>>
>> If the files are big, you could post them to our anonymous ftp site
following these instructions:
>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>
>> Thanks,
>> John
>>
>> On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>
>>> Hello MET Help,
>>>
>>> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>>>
>>> When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
>>>
>>> Thanks for your time,
>>>
>>> John D McMillen
>>>
>>>
>>> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>>> wrote:
>>>
>>>> According to our records, your request has been resolved. If you
have any
>>>> further questions or concerns, please respond to this message.
>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: JOHN DANIEL MCMILLEN
Time: Thu Mar 27 13:50:24 2014
Hey John!
Thanks for your help. I was confused, but you caught a big error
anyway.
I must have run different files through mode than the ones I sent you.
Sorry about that, but I figured out what was wrong.
When I ran mode again on the files I sent you, I got the same error
that you did. I too am running MET v4.1 and I haven't altered the
source code.
I then looked at my matlab script and found that I was dynamically
defining lat_ll and lon_ll in the NEXRAD file, but not in the WRF
file. Instead, it was fixed at what you see when you did the ncdump.
So I fixed that bug and produced new NetCDF files with identical
lat_ll, lon_ll, and grids. I then ran mode on my new files, and it
worked fine. The plots look good and I am confident the stats are
good.
Sorry for the confusion, but thank you for your help.
John
On Mar 27, 2014, at 10:24 AM, John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> John,
>
> Using ncdump, I see the following in the NetCDF files you sent:
>
> ncdump -h NEXRAD20101027.nc | egrep "lat_ll|lon_ll"
> :lat_ll = 40.1007299270073 ;
> :lon_ll = -113.69625 ;
>
> ncdump -h WRF20101027.nc | egrep "lat_ll|lon_ll"
> :lat_ll = "40.1011" ;
> :lon_ll = "-113.4123" ;
>
> As you can see, that latitudes are very close, but the longitudes
differ by .28395 degrees. That explains the 20km or so offset you're
seeing. The real mystery now is why MODE allowed you to compare
> the files in the first place!
>
> When I tried to run these two files through MODE from METv4.1, I got
the following error:
>
> ERROR : process_fcst_obs_files() -> The forecast and observation
grids do not match: Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101
lon_ll: 113.412 delta_lat: 0.009 delta_lon: 0.009 !=
> Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101 lon_ll: 113.696
delta_lat: 0.009 delta_lon: 0.009
>
> Why didn't you? What version of MET are you running? Have you made
any changes to the source code yourself?
>
> FYI, in the process of debugging your issue, I used the
plot_data_plane utility to make plots of the NEXRAD and WRF data
directly. But since it's such a small area, I decided to turn on the
plotting
> of the USA counties. Here's how:
> (1) Edit the file METv4.1/data/config/ConfigMapData
> In the "source" section, add a 4th entry for the file
"usa_county_data":
>
> source = [
> {
> file_name = "MET_BASE/data/map/legacy/world_data_minus_usa";
> },
> {
> file_name =
"MET_BASE/data/map/legacy/country_data_minus_usa";
> },
> {
> file_name = "MET_BASE/data/map/legacy/usa_state_data";
> },
> {
> file_name = "MET_BASE/data/map/legacy/usa_county_data";
> }
>
> ];
>
> Now, all of the MET tools that make plots (like MODE,
plot_data_plane, and plot_poitn
>
> (2) Then run plot_data_plane like this:
> METv4.1/bin/plot_data_plane NEXRAD20101027.nc NEXRAD20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
> METv4.1/bin/plot_data_plane WRF20101027.nc WRF20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
>
> That'll plot the data directly, each on it's own grid. And you'll
see that the resulting grids really are shifted.
>
> Hope that helps.
>
> Thanks,
> John
>
> On 03/27/2014 09:10 AM, JOHN DANIEL MCMILLEN via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>
>> John,
>>
>> Here are the files. The two .nc files are the inout files I use in
MODE. WRF is the forecast, NEXRAD is the observation. I based the
format for each on the attributes of pcp_combine NetCDF files from the
tutorial. The .ps file is the output of MODE. I believe the analysis
is accurate because the placement of the forecast and observed data is
accurate relative to one another as they are on the same grid.
Comparing the images from the .ps to the .png I created from the input
files shows the same relative placement of observation to forecast,
but shows that both fields should be over the southern portion of the
Great Salt Lake, not shifted eastward. In the .png the contour is the
same as conv_thresh in my config file. The blue contour is the
forecast and the black contour is the observation.
>>
>>
>> [cid:418d776b-a30d-422c-9d1b-daba71a8b2c0 at umail.utah.edu]
>>
>>
>> Let me know if you have any other questions.
>>
>> Thanks for your help,
>> John
>>
>> On Mar 26, 2014, at 5:15 PM, John Halley Gotway via RT
<met_help at ucar.edu>
>> wrote:
>>
>>> John,
>>>
>>> Yes, it'd help to look at some data files. You could also try
running them through the "plot_data_plane" utility, which would take
MODE out of the equation. It's likely an issue in how the grid is
>>> specified. Please send a sample NetCDF file, the PostScript
output of MODE, and a plot from Matlab showing where the data "should"
be.
>>>
>>> If the files are big, you could post them to our anonymous ftp
site following these instructions:
>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>>
>>> Thanks,
>>> John
>>>
>>> On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>>
>>>> Hello MET Help,
>>>>
>>>> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>>>>
>>>> When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
>>>>
>>>> Thanks for your time,
>>>>
>>>> John D McMillen
>>>>
>>>>
>>>> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>>>> wrote:
>>>>
>>>>> According to our records, your request has been resolved. If you
have any
>>>>> further questions or concerns, please respond to this message.
>>>>
>>>>
>>>
>>
>>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: John Halley Gotway
Time: Thu Mar 27 13:52:02 2014
John,
Great, glad you were able to iron it out. Lots of details to keep
straight.
Just let us know if any more issues or questions arise in your use of
MET and MODE.
Thanks,
John
On 03/27/2014 01:50 PM, JOHN DANIEL MCMILLEN via RT wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>
> Hey John!
>
> Thanks for your help. I was confused, but you caught a big error
anyway.
>
> I must have run different files through mode than the ones I sent
you. Sorry about that, but I figured out what was wrong.
>
> When I ran mode again on the files I sent you, I got the same error
that you did. I too am running MET v4.1 and I haven't altered the
source code.
>
> I then looked at my matlab script and found that I was dynamically
defining lat_ll and lon_ll in the NEXRAD file, but not in the WRF
file. Instead, it was fixed at what you see when you did the ncdump.
>
> So I fixed that bug and produced new NetCDF files with identical
lat_ll, lon_ll, and grids. I then ran mode on my new files, and it
worked fine. The plots look good and I am confident the stats are
good.
>
> Sorry for the confusion, but thank you for your help.
>
> John
>
>
>
>
>
> On Mar 27, 2014, at 10:24 AM, John Halley Gotway via RT
<met_help at ucar.edu>
> wrote:
>
>> John,
>>
>> Using ncdump, I see the following in the NetCDF files you sent:
>>
>> ncdump -h NEXRAD20101027.nc | egrep "lat_ll|lon_ll"
>> :lat_ll = 40.1007299270073 ;
>> :lon_ll = -113.69625 ;
>>
>> ncdump -h WRF20101027.nc | egrep "lat_ll|lon_ll"
>> :lat_ll = "40.1011" ;
>> :lon_ll = "-113.4123" ;
>>
>> As you can see, that latitudes are very close, but the longitudes
differ by .28395 degrees. That explains the 20km or so offset you're
seeing. The real mystery now is why MODE allowed you to compare
>> the files in the first place!
>>
>> When I tried to run these two files through MODE from METv4.1, I
got the following error:
>>
>> ERROR : process_fcst_obs_files() -> The forecast and observation
grids do not match: Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101
lon_ll: 113.412 delta_lat: 0.009 delta_lon: 0.009 !=
>> Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101 lon_ll: 113.696
delta_lat: 0.009 delta_lon: 0.009
>>
>> Why didn't you? What version of MET are you running? Have you
made any changes to the source code yourself?
>>
>> FYI, in the process of debugging your issue, I used the
plot_data_plane utility to make plots of the NEXRAD and WRF data
directly. But since it's such a small area, I decided to turn on the
plotting
>> of the USA counties. Here's how:
>> (1) Edit the file METv4.1/data/config/ConfigMapData
>> In the "source" section, add a 4th entry for the file
"usa_county_data":
>>
>> source = [
>> {
>> file_name =
"MET_BASE/data/map/legacy/world_data_minus_usa";
>> },
>> {
>> file_name =
"MET_BASE/data/map/legacy/country_data_minus_usa";
>> },
>> {
>> file_name = "MET_BASE/data/map/legacy/usa_state_data";
>> },
>> {
>> file_name = "MET_BASE/data/map/legacy/usa_county_data";
>> }
>>
>> ];
>>
>> Now, all of the MET tools that make plots (like MODE,
plot_data_plane, and plot_poitn
>>
>> (2) Then run plot_data_plane like this:
>> METv4.1/bin/plot_data_plane NEXRAD20101027.nc NEXRAD20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
>> METv4.1/bin/plot_data_plane WRF20101027.nc WRF20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
>>
>> That'll plot the data directly, each on it's own grid. And you'll
see that the resulting grids really are shifted.
>>
>> Hope that helps.
>>
>> Thanks,
>> John
>>
>> On 03/27/2014 09:10 AM, JOHN DANIEL MCMILLEN via RT wrote:
>>>
>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>
>>> John,
>>>
>>> Here are the files. The two .nc files are the inout files I use
in MODE. WRF is the forecast, NEXRAD is the observation. I based the
format for each on the attributes of pcp_combine NetCDF files from the
tutorial. The .ps file is the output of MODE. I believe the analysis
is accurate because the placement of the forecast and observed data is
accurate relative to one another as they are on the same grid.
Comparing the images from the .ps to the .png I created from the input
files shows the same relative placement of observation to forecast,
but shows that both fields should be over the southern portion of the
Great Salt Lake, not shifted eastward. In the .png the contour is the
same as conv_thresh in my config file. The blue contour is the
forecast and the black contour is the observation.
>>>
>>>
>>> [cid:418d776b-a30d-422c-9d1b-daba71a8b2c0 at umail.utah.edu]
>>>
>>>
>>> Let me know if you have any other questions.
>>>
>>> Thanks for your help,
>>> John
>>>
>>> On Mar 26, 2014, at 5:15 PM, John Halley Gotway via RT
<met_help at ucar.edu>
>>> wrote:
>>>
>>>> John,
>>>>
>>>> Yes, it'd help to look at some data files. You could also try
running them through the "plot_data_plane" utility, which would take
MODE out of the equation. It's likely an issue in how the grid is
>>>> specified. Please send a sample NetCDF file, the PostScript
output of MODE, and a plot from Matlab showing where the data "should"
be.
>>>>
>>>> If the files are big, you could post them to our anonymous ftp
site following these instructions:
>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>>>
>>>> Thanks,
>>>> John
>>>>
>>>> On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>>>>
>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>>>
>>>>> Hello MET Help,
>>>>>
>>>>> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>>>>>
>>>>> When I plot the data with matlab I do not see the same placement
error. Is there something I have incorrectly specified in the config
file perhaps? Or did I incorrectly spoof an attribute that tells MET
to use a latlon grid vice some other grid? I can provide files if
required.
>>>>>
>>>>> Thanks for your time,
>>>>>
>>>>> John D McMillen
>>>>>
>>>>>
>>>>> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>>>>> wrote:
>>>>>
>>>>>> According to our records, your request has been resolved. If
you have any
>>>>>> further questions or concerns, please respond to this message.
>>>>>
>>>>>
>>>>
>>>
>>>
>>
>
>
------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #65939] Resolved: Help
From: JOHN DANIEL MCMILLEN
Time: Thu Mar 27 14:02:14 2014
John,
Will do. Thanks again.
John
On Mar 27, 2014, at 1:52 PM, John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> John,
>
> Great, glad you were able to iron it out. Lots of details to keep
straight.
>
> Just let us know if any more issues or questions arise in your use
of MET and MODE.
>
> Thanks,
> John
>
> On 03/27/2014 01:50 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>
>> Hey John!
>>
>> Thanks for your help. I was confused, but you caught a big error
anyway.
>>
>> I must have run different files through mode than the ones I sent
you. Sorry about that, but I figured out what was wrong.
>>
>> When I ran mode again on the files I sent you, I got the same error
that you did. I too am running MET v4.1 and I haven't altered the
source code.
>>
>> I then looked at my matlab script and found that I was dynamically
defining lat_ll and lon_ll in the NEXRAD file, but not in the WRF
file. Instead, it was fixed at what you see when you did the ncdump.
>>
>> So I fixed that bug and produced new NetCDF files with identical
lat_ll, lon_ll, and grids. I then ran mode on my new files, and it
worked fine. The plots look good and I am confident the stats are
good.
>>
>> Sorry for the confusion, but thank you for your help.
>>
>> John
>>
>>
>>
>>
>>
>> On Mar 27, 2014, at 10:24 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>> wrote:
>>
>>> John,
>>>
>>> Using ncdump, I see the following in the NetCDF files you sent:
>>>
>>> ncdump -h NEXRAD20101027.nc | egrep "lat_ll|lon_ll"
>>> :lat_ll = 40.1007299270073 ;
>>> :lon_ll = -113.69625 ;
>>>
>>> ncdump -h WRF20101027.nc | egrep "lat_ll|lon_ll"
>>> :lat_ll = "40.1011" ;
>>> :lon_ll = "-113.4123" ;
>>>
>>> As you can see, that latitudes are very close, but the longitudes
differ by .28395 degrees. That explains the 20km or so offset you're
seeing. The real mystery now is why MODE allowed you to compare
>>> the files in the first place!
>>>
>>> When I tried to run these two files through MODE from METv4.1, I
got the following error:
>>>
>>> ERROR : process_fcst_obs_files() -> The forecast and observation
grids do not match: Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101
lon_ll: 113.412 delta_lat: 0.009 delta_lon: 0.009 !=
>>> Projection: Lat/Lon Nx: 274 Ny: 229 lat_ll: 40.101 lon_ll: 113.696
delta_lat: 0.009 delta_lon: 0.009
>>>
>>> Why didn't you? What version of MET are you running? Have you
made any changes to the source code yourself?
>>>
>>> FYI, in the process of debugging your issue, I used the
plot_data_plane utility to make plots of the NEXRAD and WRF data
directly. But since it's such a small area, I decided to turn on the
plotting
>>> of the USA counties. Here's how:
>>> (1) Edit the file METv4.1/data/config/ConfigMapData
>>> In the "source" section, add a 4th entry for the file
"usa_county_data":
>>>
>>> source = [
>>> {
>>> file_name =
"MET_BASE/data/map/legacy/world_data_minus_usa";
>>> },
>>> {
>>> file_name =
"MET_BASE/data/map/legacy/country_data_minus_usa";
>>> },
>>> {
>>> file_name = "MET_BASE/data/map/legacy/usa_state_data";
>>> },
>>> {
>>> file_name = "MET_BASE/data/map/legacy/usa_county_data";
>>> }
>>>
>>> ];
>>>
>>> Now, all of the MET tools that make plots (like MODE,
plot_data_plane, and plot_poitn
>>>
>>> (2) Then run plot_data_plane like this:
>>> METv4.1/bin/plot_data_plane NEXRAD20101027.nc NEXRAD20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
>>> METv4.1/bin/plot_data_plane WRF20101027.nc WRF20101027.ps
'name="APCP"; level="(*,*)";' -plot_range 0 50
>>>
>>> That'll plot the data directly, each on it's own grid. And you'll
see that the resulting grids really are shifted.
>>>
>>> Hope that helps.
>>>
>>> Thanks,
>>> John
>>>
>>> On 03/27/2014 09:10 AM, JOHN DANIEL MCMILLEN via RT wrote:
>>>>
>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>>
>>>> John,
>>>>
>>>> Here are the files. The two .nc files are the inout files I use
in MODE. WRF is the forecast, NEXRAD is the observation. I based the
format for each on the attributes of pcp_combine NetCDF files from the
tutorial. The .ps file is the output of MODE. I believe the analysis
is accurate because the placement of the forecast and observed data is
accurate relative to one another as they are on the same grid.
Comparing the images from the .ps to the .png I created from the input
files shows the same relative placement of observation to forecast,
but shows that both fields should be over the southern portion of the
Great Salt Lake, not shifted eastward. In the .png the contour is the
same as conv_thresh in my config file. The blue contour is the
forecast and the black contour is the observation.
>>>>
>>>>
>>>> [cid:418d776b-a30d-422c-9d1b-daba71a8b2c0 at umail.utah.edu]
>>>>
>>>>
>>>> Let me know if you have any other questions.
>>>>
>>>> Thanks for your help,
>>>> John
>>>>
>>>> On Mar 26, 2014, at 5:15 PM, John Halley Gotway via RT
<met_help at ucar.edu>
>>>> wrote:
>>>>
>>>>> John,
>>>>>
>>>>> Yes, it'd help to look at some data files. You could also try
running them through the "plot_data_plane" utility, which would take
MODE out of the equation. It's likely an issue in how the grid is
>>>>> specified. Please send a sample NetCDF file, the PostScript
output of MODE, and a plot from Matlab showing where the data "should"
be.
>>>>>
>>>>> If the files are big, you could post them to our anonymous ftp
site following these instructions:
>>>>> http://www.dtcenter.org/met/users/support/met_help.php#ftp
>>>>>
>>>>> Thanks,
>>>>> John
>>>>>
>>>>> On 03/26/2014 03:21 PM, JOHN DANIEL MCMILLEN via RT wrote:
>>>>>>
>>>>>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=65939 >
>>>>>>
>>>>>> Hello MET Help,
>>>>>>
>>>>>> I am using mode to analyze NEXRAD accumulations and WRF
accumulations that I am formatting as pcp_combine netcdf files. I am
using the NEXRAD latlon grid as the common grid and have filled out
the netcdf attributes as such. The tool is working and the analysis
seems to be correct. The issue I have is the data is not plotted
correctly with respect to the maps in the .ps file output by mode. It
appears that the data is being plotted ~20km east of the plotted map.
>>>>>>
>>>>>> When I plot the data with matlab I do not see the same
placement error. Is there something I have incorrectly specified in
the config file perhaps? Or did I incorrectly spoof an attribute that
tells MET to use a latlon grid vice some other grid? I can provide
files if required.
>>>>>>
>>>>>> Thanks for your time,
>>>>>>
>>>>>> John D McMillen
>>>>>>
>>>>>>
>>>>>> On Mar 24, 2014, at 9:58 AM, John Halley Gotway via RT
<met_help at ucar.edu>
>>>>>> wrote:
>>>>>>
>>>>>>> According to our records, your request has been resolved. If
you have any
>>>>>>> further questions or concerns, please respond to this message.
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>>
>
------------------------------------------------
More information about the Met_help
mailing list