Alternative to bivariate interpolation from akima-package

Dear list,

A function in the intamap package currently uses the bivariate 
interpolation method for irregularly spaced data from the akima package. 
This dependency is a problem because of license mismatch (intamap is GPL 
 >=2, akima has an exclusively non-commercial license). Does anyone know 
about a good alternative to this method?
The method must be able to interpolate from an irregularly spaced 
observation network to a grid, and the interpolated surface must be 
smooth, i.e., differentiable. The method should ideally be fast also for 
larger data sets. The thin plate splines regression in the fields 
package (Tps) seems to produce a smooth surface, but struggles with 
memory consumption when the observation network is large (>2000 points).
mba.surf in the MBA package is very fast, but it could run into problems 
with heterogeneous sampling patterns.

Any other suggestions? Is there a natural neighbor interpolation method 
implemented in R?

Best wishes,
Jon

Jon Olav Sk?ien
Joint Research Centre - European Commission
Institute for Environment and Sustainability (IES)
Land Resource Management Unit

Via Fermi 2749, TP 440,  I-21027 Ispra (VA), ITALY

Disclaimer: Views expressed in this email are those of the individual and do not necessarily represent official views of the European Commission.