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

[Ryo Furue <furue@xxxxxxxxxx>]

Hi Steve,

| 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. [. . .]  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.

Aha!  Thanks.  That explains the result of my first experiment:

    fill/. . ./hlimits=10:30 rose

Since ferret assumes the longitude axis is cyclic, in

   fill/. . ./hlimits=10:30/x=10:30 rose

the 10 <= x <= 20 portion of the data is supplied
from 370 <= x <= 380.  That was my second experiment.

So, the problem which occurs with

   fill/. . ./hlimits=10:30/x=9:31 rose

(my third exp.) may be related to the cyclicity.

To summarize the remaining questions/problems:

1) "hlimits=10:30/x=9:31" should work even when there's
   a "branch cut" within the range. (Compare the 3rd plot
   with the 2nd in my webpage
   http://iprc.soest.hawaii.edu/~furue/ferret-contours/index.html)

2) Shouldn't "hlimits=10:30" imply "x=10:30" when the user
   omits "x=" ?  (It doesn't in the current implementation,
   as Steve explains.)

Cheers,
Ryo