Re: Question of natgrid

Hi users,
  Thanks for your suggestion. I have two question of the variable in
'linint2_points'. First, it said x and y are a strictly monotonically
increasing array. Does it mean than I need to rearrange my data by
monotonically increasing first?
  Second, my fi is the value I want to interpolate. I don't understand what
it means by that the two rightmost dimensions (*nyi* x *nxi*) are the
dimensions to be used in the interpolation. Could you explain me how I
should revise my fi to do the linint2_points. Thank you.

On Mon, May 16, 2011 at 8:44 AM, <debasish@sca.uqam.ca> wrote:

> "natgrid" interpolates from 2D random data to get values on a rectangular
> output grid, therefore, if you like to interpolate to a grid with
> lat(nlat) and lon(mlon), then the output dimension should be nlatxmlon.
> But I believe you like to interpolate to arbitrarily specified coordinate
> pairs instead of rectangular output grid. Please try "linint2_points"
>
> Also check
> http://www.ncl.ucar.edu/Document/Functions/interp.shtml
> http://www.ncl.ucar.edu/Document/Functions/regrid.shtml
>
>
> Thanks
>
> Debasish
>
>
> > Hi users,
> > I have a question about natgrid interpolation. When I use natgrid to
> > interpolate, I give 100 new coordinates to interpolate. However, I got
> > 100X100 new value after interpolation. Could anyone please explain me why
> > I
> > get 10000 new values instead of 100 new values? My goal is interpolate
> 161
> > value with lat and lon to 45320 new coordinates. When I try to use 45320
> > new
> > coordinates in the script, the program will crash because of huge
> > calculation (45320x45320?). My new coordinates is irregularly spaced
> grid.
> > Thus, I have tried dspnt2 for interpolation, but the result is not
> > satisfying. Except dspnt2 and natgrid, could anyone please give me some
> > advice to process the interpolation? Thank you.
> >
> > --
> > Best regards
> > Wei-Chih
> > _______________________________________________
> > ncl-talk mailing list
> > List instructions, subscriber options, unsubscribe:
> > http://mailman.ucar.edu/mailman/listinfo/ncl-talk
> >
>
>
>

-- 
Best regards
Wei-Chih
-- 
Best regards
Wei-Chih

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon May 16 12:16:42 2011