Fig. 4

Canonical correspondence analysis (CCA) plot, showing the correspondence (~influence) of main environmental factors (blue arrows; cf. Table 4) with the vegetation composition of the plots (black crosses; vegetation types R1, R2, B1 and B2 of the plots are indicated). The analysis was based on the 30 most important trees species (each one shown as a red dot). The analysis was significant overall at p < 0.001 (permutation test for CCA). The eigengrad values of the environmental variables (providing a measure of importance of each variable to explain gradients in vegetation patterns, listed from the strongest to the weakest) were 0.119 for ‘elevation’, 0.109 for ‘creek’, 0.069 for ‘tree density’, 0.054 for ‘slope’, 0.038 for ‘dead trees (%)’, 0.035 for ‘soil OM’, 0.034 for ‘dead TeAGB (%)’, 0.031 for ‘sapling density’, 0.029 for ‘TeAGB’, 0.029 for ‘soil pH’, 0.027 for ‘soil N’, and 0.022 for ‘soil water’