interpolation question:

Hi Joel,

I have a question about interpolation:

I'm trying to replace missing values in data to be converted for hysplit
- they seem to be entirely due to interpolating model output to standard
pressure levels and the fact that the lower levels dip below the surface
and thus become undefined.

I was hoping to be able to do a multilinear/multidimensional
interpolation, but the most ncl will do is bilinear (ignoring time and
pressure variations).

I read online that piecewise linear interpolation is a more efficient
form of multilinear interpolation, but it seems to be less accurate than
a multidimensional spline would be.

Do you know much about this, and would you recommend one over the other?
For the time being I'm going with the latter and using the linmsg
function in ncl.

Thanks,
guillaume

-- 
guillaume mauger
http://ieng9.ucsd.edu/~gmauger/
_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk