The aim of this paper is to identify the main factors that determine territorial differences between two large neighbouring provinces (Granada and Almería) that each occupy both sides of the Sierra Nevada mountain range in southern Spain. Our analysis produced a prediction model for riverside woodland cover, one of the variables that most strongly characterizes the differences between the two provinces. A general linear model was constructed for spatial analysis (spherical model with REML estimation), taking into account the location of the municipalities (longitude, latitude, and elevation) and other factors related to soil type, vegetation, or woodland cover. Interpolation of spatial data was performed by ordinary kriging. The predictions obtained over a grid of locations enabled patterns in each region to be derived and visualised, with very clear results. The pattern of riverside woodland cover is highly differentiated between the two provinces, despite their presenting adjoining areas.