Re: [ferret_users] contours, hlimits=, x=, ?

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi Ryo,
So, what you did at the end of your script SHOULD have worked, 
specifying a larger region than desired with the /X= and /Y= qualifiers 
to fill to the edges of the /HLIMITS/VLIMITS region.  But the fill 
doesn't extend to the edges in Y.

   yes? fill/hlimits=10:30/vlimits=80s:60s/x=9:31/y=81s:59s rose

We'll look into why this happens.  But notice that the following command 
does give you th plot, completely filled in to the edges:

   yes? fill/x=10:30/y=80s:60s rose

Ansley

Steve Hankin wrote:

> Hi Ryo,
>
> X= specifies the data range
> HLIMITS= specifies the clipping region of the plot axes
>
> The etopo20 dataset is defined on a longitude axis that runs from 20 
> to 380, so that its "branch point" does not disect a major ocean 
> basin.  (Try "SHOW GRID rose")  Your graphical axis limits 
> "hlimits=10:30" span this branch cut.  Since your FILL command does 
> not specify the coordinate range explicitly (no "/X=" is specified) 
> the Ferret IO layer supplies to the graphics layer the full native 
> coordinate range, which is X=21:379.  Thus your specified axis range 
> of 10:30 overlaps the data for only half of its range, 21:30.  So half 
> of your plot is blank.
>
>    - steve
>
> Ryo Furue wrote:
>
> > Hi Ferret users and developers,
> >
> > I reported a similar problem a while ago, but now things
> > are looking "curiouser and curiouser".  Basically, I'm wondering
> > what happens and what should happen when hlimits= (vlimits=) is
> > specified for a contour plot.
> >
> > Here's little experiment I've done:
> >
> >  http://iprc.soest.hawaii.edu/~furue/ferret-contours/index.html
> >
> > First, when I don't specify x= or y=,
> >
> >  yes? set data etopo120
> >  yes? fill/lev=(-6000,4000,1000,-1)/hlimits=10:30/vlimits=80s:60s \
> >       rose
> >
> > I get a huge white region, as seen in the plot in my webpage cited
> > above.  Is this just me? or does this happen to other users?
> >
> > Next, when I specify both hlimits (vlimits) and x= (y=),
> >
> >  yes? fill/. . ./hlimits=10:30/vlimits=80s:60s \
> >           /x=10:30/y=80s:60s   rose
> >
> > I get a clean plot.
> >
> > The plot isn't quite satisfactory, though, because there is a white
> > region along the edges (and I think I know why).
> > Since we have datapoints outside our plotting region,
> > why not use them?
> >
> >  yes? fill/. . ./hlimits=10:30/vlimits=80s:60s \
> >           /x=9:31/y=81s:59s rose
> >
> > But, this plan partially fails.  When we extend only in the y-direction
> > (y=81s:59s), the contouring region is extended to the northern and 
> > southern edges.  Good.  But, when we extend in the
> > x-direction (x=9:31), we get a weired plot, as shown in my webpage.
> >
> > Maybe am I doing something which shouldn't work?
> >
> > Regards,
> > Ryo
> >