Re: Poisson_grid_fill

I am not sure what you mean by "random grid data" or
"works physically".

[1] poisson_grid_fill does not work for *truly* random data.
      If you have truly random data, you could assign each
      value to the nearest grid point. Then use poisson_grid_fill.

See: http://www.cgd.ucar.edu/~shea/poisson_grid_fill.png

[2] If there are gaps [missing values indicated by _FillValue] in
     a grid then it will use the existing valid values as boundary values.
     Existing grid values are not altered.

     Note that poisson_grid_file will extrapolate. However, extrapolation
     is always dangerous!

HTH

Roberta Perkins wrote:
> Hello,
>
> I'm trying to figure out how Poisson_grid_fill works physically. Are there any
> references that look at solving Poisson's equation using relaxation for random
> grid data with missing values?
>
> Sincerely,
> Roberta Perkins
>

_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Feb 18 2008 - 19:05:55 MST