[Met_help] strange black background for MODE .ps files

[John Halley Gotway]

Shelley,

Our plan was to add an "output_prefix" parameter to the MODE configuration file.  So it wouldn't be on the command line.

However, since you can use environment variables in the config files, I'd suggest setting the output_prefix to be equal to some environment variable in the config file.  Then reset that environment
variable to what you'd like on the command line before calling MODE.

I believe that would accomplish what you're trying to do.

Sound reasonable?

John

Holmberg, Shelley wrote:
> Hi John,
> 
> Thank you, that is very helpful!  The wrf_precip2.ctable is more clear than the mode_raw.ctable, especially since it de-emphasizes the the lesser precipitation accumulations.  The user-defined naming would be very helpful, especially in the mode command line where it would be easy to script a loop.
> 
> Thank you again!
> Shelley
> 
> 
> 
> 
> -----Original Message-----
> From: John Halley Gotway [mailto:johnhg at rap.ucar.edu]
> Sent: Fri 3/20/2009 5:05 PM
> To: Holmberg, Shelley
> Cc: met_help at ucar.edu
> Subject: Re: [Met_help] strange black background for MODE .ps files
>  
> Shelley,
> 
> Thanks for sending the postscript files.  I took a look at them, and I
> don't see a problem there.  I believe you just have a lot of data in there
> that's around 0 - which is a very dark green in the default color table.
> 
> I'd suggest changing the color table used to generate the plots to be one
> that you like better.  Try editing your MODE config files as follows:
> - In the settings for "fcst_raw_color_table" and "obs_raw_color_table",
> change "mode_raw.ctable" to "wrf_precip2.ctable".
> - Set "fcst_raw_plot_min" and "obs_raw_plot_min" to 0.0.
> - Set "fcst_raw_plot_max" and "obs_raw_plot_max" to a reasonably large
> value.  Any values greater than this max will be plotted using the color
> for the maximum value.
> 
> You're welcome to try some of the other colortables or create your own if
> you like.  I'm suggesting that you set the plot min and max values so that
> the colorscale is always fixed.  That way when you compare plots from
> different models or times the colors in the plots will correspond to the
> same values.  Otherwise, the raw colortable is scaled to the actual
> min/max values in the field.
> 
> In the upcoming release of MET (in early April), I'll change the default
> MODE config files to use the "wrf_precip2.ctable" colortable for
> precipitation.
> 
> As for the output file names from MODE, it sounds like you'd like to
> include the "MODEL" name in the output file names.  Would that solve your
> issues?  That change would require changing the source code and
> recompiling, but I think it's a good idea.  I'll talk to the others in our
> development team and see if they'd like to make that change.
> 
> Perhaps it'd be even better to allow the user to specify a file prefix
> name in the MODE config file.  Currently, that would just be "mode_", but
> then the user could change it to "mode_${MODEL}".  How does that sound to
> you?
> 
> Thanks,
> John
> 
>> Hi!
>>
>> Most of my MODE .ps files have this strange blacked out background while
>> some have the normal white background with the states.  I am using the
>> exact same configuration file for all and I have scripted the WRF
>> postprocessing and pcp_combine steps to be the exact same for all of my
>> model runs.  Attached are two .ps files from MODE, they are of two
>> different model runs (enkf0 and enkf1) valid at the same date and time
>> with the exact same observations.  Have you seen this before?  Could it be
>> a thresholding issue?
>>
>> Also, is there a way to rename the output files in the ./mode step?  I
>> have been using -outdir and then changing the name manually but it is
>> tedious and fraught with peril as I am analyzing six model versions and
>> they are all named the same thing just in different directories.
>>
>> Thank you so much!
>> Shelley Holmberg
>> _______________________________________________
>> Met_help mailing list
>> Met_help at mailman.ucar.edu
>> http://mailman.ucar.edu/mailman/listinfo/met_help
>>
> 
> 
> 
>