Modelling Resistance Traits Across A Landscape

Modelling resistance traits across a landscape using different interpolation methods

[2019-12-02]

Simple linear interpolation via interp (akima package) was performed and produced the plot on the left with yellow (low resistance) to red (high resistance) heatmap.

Smooth interpolation via generalized additive modelling: gam (gmcv package) was also performed and produced the graph on the bottom right. However do the spline- and tensor-based GAM interpolation need to account for physical boundaries to get more sensible resistance gradient plots? And can we acquire more data points to improve the gradient map?