trend surface analysis

Hello,

What is the better way to perform trend surface analysis in R (with GRASS data),
with third-order surfaces (or higher levels?)

Almost certainly surf.ls() in the spatial package, because it transforms 
the coordinates internally to lie within -1, +1. So numerically it will be 
more stable than say higher powers of numbers of metres south or north of 
the Equator in UTM. Some thought is need to get the points of the 
prediction grid right for GRASS, though; see the trmat.G() wrapper 
function.

Roger

thanks

Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Breiviksveien 40, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93
e-mail: Roger.Bivand at nhh.no