Comparison of Three Supervised Learning Methods for Digital Soil Mapping: Application to a Complex Terrain in the Ecuadorian Andes

Figure 9

Soil texture maps predicted by gradient boosting with overlaid hillshading from north, (a) sand content, (b) silt content, and (c) clay content and marginal dependence plots displaying the importance of predictors for the sand content (d) and the clay content (e).

(a)  

(b)

(c)

(d)

(e)