Hi Rich,
When you ran your script, did you get a bunch of warnings:
warning:ComputeMinorTickMarks: The distance between major tickmarks is so close or so far apart that arithmetic error may cause problems determining the accurate locations of minor tickmarks
This is because the "time" array on your X axis has values like 2444050, 2444051,…,2444074
I don't know if this is a bug that AreaFill can't handle this kind of spacing (I will file a ticket on this), but you can fix this with:
;---Fix the time values so they're not so large.
var&time = var&time - 2444000
Once you do this, the plots look fine (see attached).
This is just a kludgy work-around. You may want to check the "cd_convert" function for converting your time values to a more reasonable set of values, but you may have to fix the "calendar" and "units" of var&time before you can use this function.
--Mary
On May 21, 2014, at 5:38 PM, Rich Neale <rneale@ucar.edu> wrote:
> Hi,
>
> I have a bug when using cnFillMode = "AreaFill" for a simple contour plot (gsn_csm_hov).
> 2 things happen: 1. Contour intervals are randomly absent and 2. Contour intervals are missed on their way to max and mins.
> The plot attached shows the problem. The left is with AreaFIll and the right is with RasterFill which seems accurate.
> I am using ncl version
> /glade/u/apps/opt/ncl/6.2.1/gnu/4.4.6/bin/ncl
> And on yellowstone (yslogin6)
> My script is attached and it points to glade data.
>
> Many thanks!
> Rich
>
>
> <contour_bug.ps><ncl-talk-eg.ncl>_______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu May 22 15:47:58 2014
This archive was generated by hypermail 2.1.8 : Tue May 27 2014 - 15:45:08 MDT