[ferret_users] accuracy of axis in ferret

[Martin Schmidt <martin.schmidt@xxxxxxxxxxxxxxxxx>]

Hi,
I am trying to read some large part of the GEBCO topography with ferret. 
The file does not have axes as usual but start-end-space definition.

With ncdump one sees 

x_range = -60, 50 ;

y_range = -50, 20 ;

z_range = -7861, 5778 ;

spacing = 0.00833333333333333, 0.00833333333333333 ;

dimension = 13200, 8400 ;

So I try to set up a grid from this definition:

define axis/x=`X_RANGE[i=2]`:`X_RANGE[i=1]`:`SPACING[i=2]`/edges 
/units="degrees_east" xax
define axis/y=`Y_RANGE[i=2]`:`Y_RANGE[i=1]`:`SPACING[i=1]`/edges 
/units="degrees_north" yax
define grid/x=xax/y=yax xygrid

Now to see the result:
say `x[gx=xygrid],return=iend`
!-> MESSAGE/CONTINUE 13201
13201

Hence the resulting axis has wrong size. The same happens for the 
y-axis. Mapping topography to this grid gives a strange looking piece of 
the earth.

However, the varables x_range and so on are defined as doubles. Why is 
this ignored by ferret?

Playing around with the axis definition gives better results:
define axis/x=`X_RANGE[i=1]`:`X_RANGE[i=2]`:`0.0083334`/edges 
/units="degrees_east" xax
say `x[gx=xygrid],return=iend`
!-> MESSAGE/CONTINUE 13200
13200

However, I want to process several pieces of the topography and I would 
prefer a clean solution: reading data from file and writing the results 
without any adjustment of scripts.
Is this possible? What I am doing wrong here?

Greetings,
Martin