A multi-data ensemble approach for predicting woodland type distribution: Oak woodland in Britain.

A Kolmogorov–Smirnov test showed a highly significant difference ( p ‐value ≤ .001, D ‐value 0.28) between the biomod2 oak probability rasters “filtered by area” and “filtered by area and elevation.” Figure 4 shows that forcing pixels into the elevation distribution of the NFI survey sample squares caused a low probability tail in the distribution compared with the area‐only filter.

also in 132,142 other papers