[Met_help] Plotting VSDB files.

[Lacey Holland]

Hi, Brian!


FYI, I'm cc'ing met_help on this, as I think it may be of interest to a 
few others in our group.


That's great you are able to use awk and gnuplot!  Those are very 
reasonable alternatives.  We have considered providing some scripts for 
those, but they haven't seemed to be as in high demand as some of the 
other packages.  It's good to know you are using them, though, and that 
they work.  The beta release was somewhat limiting on some of the 
statistics that it produced, but the next release will provide a VSDB 
analysis tool that you should also be able to use with awk and gnuplot.  
The VSDB analysis tool will compute aggregate statistics over whatever 
cases you provide VSDB files.  From the output of that analysis tool, 
you'll be able to do a lot more with gnuplot, such as plot the median 
absolute error vs. height for your model runs.


We are also starting to compile example scripts/programs for reading the 
VSDB and other output files that we might show as examples on our 
website.  If you feel comfortable doing it, it might be really helpful 
to other users if you could send over some of your awk and gnuplot code 
so that we could show it as an example of how someone might use them.


The library linking problem you mentioned, -lfg2c vs. -lf2c, is one that 
others have had as well.  That will be better documented in the next 
release, but thanks for letting us know!  Please let us know if you run 
into anything else that doesn't work as documented. 


Thanks again and take care,
--Lacey Holland



Brian Gaudet wrote:
>
> Hi Lacey,
>
>
> Yes, that makes sense from what I know, but I personally do not have 
> experience
> with IDL.  Others here I'm sure do.  I've actually generated plots 
> using the
> 'awk' utility and 'gnuplot' on the .txt output, which is rather basic 
> but gets the
> job done because as you say they are ASCII.
>
> I'm glad everything seems to be working, but we haven't used 'grid_stat'
> or the 'mode' package yet.  One minor change we had to perform to get
> pb2nc working on our Linux system was change a library link in the 
> compiler
> options from '-lf2c' to '-lg2c', since the two are not equivalent 
> here, and
> the compiler used is ultimately g77.
>
> Brian
>
>
>
>
>
>
> On Mon, Nov 12, 2007 05:27 PM, *Lacey Holland <lholland at ucar.edu>* wrote:
>
>     Hi, Brian!
>
>     Thanks for inquiring about the VSDB files.  We're certainly planning to 
>     provide subroutines/scripts in a few languages (IDL, R, NCL, etc.)
>     for 
>     plotting VSDB files but unfortunately, those aren't available yet. 
>
>     In the formal release that is coming up soon, there will be analysis 
>     tools available for sorting through VSDB files and calculating other 
>     quantities from the VSDB files, but these aren't available right now in 
>     the Beta version.  It will likely be sometime after the formal release 
>     when we provide the subroutines mentioned in the User's Guide for plotting.
>
>     I have had a great deal of success on my own plotting the VSDB output in 
>     IDL, however.  I think I have seen others use R.   Right now, these 
>     subroutines aren't formally available, though.  I have found the VSDB 
>     files aren't difficult to work with in IDL, especially if you "cat"
>     the 
>     files together before reading them in.  Since they are in ASCII format, 
>     most software packages should be able to read them. 
>
>     Anyhow, thanks so much for your interest, and it's exciting to know you 
>     were able to get that sort of output!  Please let us know if there's 
>     anything else we can help with.
>
>
>     Thanks again and take care,
>     --Lacey Holland
>
>
>
>
>
>     Barbara Brown wrote:
>     > Dear Brian,
>     >
>     > All of the MET team members are attending a retreat for the NCAR 
>     > Research Applications Laboratory this week. Someone from the team will 
>     > get in touch with you next week.
>     >
>     > It's great to know that you've gotten some results from point_stat!
>     >
>     > Best regards,
>     >
>     > Barbara Brown
>     >
>     > Brian Gaudet wrote:
>     >
>     >
>     > Hello,
>     >
>     >
>     > I am Brian Gaudet, a research associate working under
>     > Dave Stauffer and Nelson Seaman at PSU.
>     > We are interested in using the MET beta package
>     > to verify some of our realtime simulations
>     > of WRF-ARW.
>     >
>     > We have succeeded in comparing our results
>     > with prepbufr files using point_stat, and
>     > generating vsdb (and general text) output.
>     > We were wondering if you have any suggested
>     > software packages for generating graphical
>     > output from these files, and/or useful scripts
>     > for doing so.  (I see you mention IDL in the
>     > users guide).
>     >
>     >
>     > Thank you for any help you can provide.
>     >
>     >
>     > Sincerely,
>     >
>     > Brian Gaudet
>     > bjg20 at psu.edu
>     >
>     > ------------------------------------------------------------------------
>     >
>     > _______________________________________________
>     > Met_help mailing list
>     > Met_help at mailman.ucar.edu
>     > http://mailman.ucar.edu/mailman/listinfo/met_help
>     >   
>     >
>
>
>     -- 
>     ____________________________
>     Lacey Holland
>
>     UCAR/NCAR, Research Applications Laboratory
>     P.O. Box 3000
>     Boulder, CO  80307
>     (303) 497-8396
>
>
>
>         
>


-- 
____________________________
Lacey Holland

UCAR/NCAR, Research Applications Laboratory
P.O. Box 3000
Boulder, CO  80307
(303) 497-8396