Re: [ferret_users] plotting grid point density

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

Hi Mark,
I think this is just the resolution of the screen or output file that's causing the bands of missing lines. If you make the plot, and then say

  yes? SHOW SYM PPL$XPIXEL
  yes? SHOW SYM PPL$YPIXEL

On my system, this is 867 by 748 pixels so it's just a couple of pixels per degree longitude, and in fact each grid cell is some non-integer number of pixels wide. Try your plot on a sub-region of the globe and you'll see there are no missing lines.

Another way to visualize the grid density comes from a script from Andrew Wittenberg a while back.  Define a "checkerboard" variable

   yes? let checkerboard = 0*temp + mod(i+j,2)
   yes? shade/k=1/palette=greyscale/lev=(0,1,1)/nokey  checkeerboard

You'll see some of the same pixel-resolution artifacts here. A sub-region works great.

Ansley


Mark.Collier@csiro.au wrote:

I am trying to display horizontal plots/depth sections grid density (without thinning) using the following approach (I’m just using levitus data here as an eg):

 

use "/cs/datastore/csdar/cec006/ferret/data/levitus_climatology.cdf"

 

let a=i+temp[d=1,k=1,l=1]-temp[d=1,k=1,l=1]

let b=j+temp[d=1,k=1,l=1]-temp[d=1,k=1,l=1]

 

cont/size=0.0001/lev=(1,360,1) a

cont/size=0.0001/ove/lev=(1,180,1) b

 

let a=k+missing(temp[d=1,i=1,l=1],0)-missing(temp[d=1,i=1,l=1],0)

let b=j+missing(temp[d=1,i=1,l=1],0)-missing(temp[d=1,i=1,l=1],0)

 

cont/size=0.0001/lev=(1,180,1) b

cont/size=0.1/ove/lev=(1,20,1) a

 

I find that sometimes there are white bands of missing lines but don’t know where they are coming from. In the case of the section I would also like to see the 20^th line but it only goes as far as 19.

 

There may be some other way of showing data density, which I’d be happy to hear about. I tried plotting symbols using plot/vs but found that it created some strange looking patterns, particularly if the grid point density is high.

 

Regards,

Mark.