Date: Sun, 28 Sep 2008 21:11:44 -0400
Hello,
I have come across some "weirdness" in NCL contouring that I have not
been able to resolve and was hoping for some help. The attached figures
show what I hope to produce and what I'm getting for specific datasets
that have _FillValue data in them. In the "bad_plot" image, the central
unfilled area is being plotted correctly as it represents _FillValue
data. The areas "above" and "below" are also handled correctly. It's
only the areas to each side that go crazy with contours (and I get the
same result without the color fill).
I am building these netCDF files myself using Python/pycdf (script
cod_to_nc, in the tar file) and have not been able to find any oddities
in these files that would explain the contouring madness. I have also
looked at the data extensively as the NCL script is running with no
insights. At this point it seems like something amiss in NCL -- though
this is my first real experience using files with FillValues so I can't
rule out entirely that it's my fault.
The link below is to a tgz-file with the script, data and larger (4x3
panels) examples of my output. The script (plot_var.ncl) takes a netCDF
file as input and produces a postscript output. To reproduce the files
that have these sample images, use these commands respectively:
ncl 'var="h500z"' plot_var.ncl
ncl 'var="h700z_msk"' plot_var.ncl
I am using NCL 5.0.0 and seeing the same results on both Mac OS 10.5.5
and 10.4.11.
Script, data, output here: http://www.geosc.psu.edu/fetch.php?file=976
Thanks for your time,
Dave Reusch
-- 517 Deike Building Earth & Environmental Systems Institute Penn State University University Park, PA 16802
_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
