*3.3. Relative Importance of Leaf Age, Tree Age, Tree Height as well as Leaf Nutrients on the* δ*13C and* δ*15N Values*

In this analysis, we assumed that the effect of leaf age may not be linear. Among the studied variables, leaf age, tree height and tree age have accounted for 18.78% variance in the model (Figure 4A). The independent effects of leaf age showed a larger contribution (*R*<sup>2</sup> = 14.49%, *p* < 0.05) to total variation in leaf δ13C compared with tree age and tree height. In addition, leaf nutrients such as leaf nitrogen and phosphorus concentrations were the other most important predictors for the δ13C (*R*<sup>2</sup> = 33.56%). In particular, leaf N concentrations explained the largest percentage of variation in leaf δ13C and its effect was significant (*R*<sup>2</sup> = 19.24%, *p* < 0.01).

Leaf physiological properties (leaf age, tree height and tree age) and leaf nutrients (N and P concentrations) together explained 47.29% of the variations in leaf δ15N (Figure 4B). Among the study variables, leaf age was the most important predictor for the δ15N (*R*<sup>2</sup> = 15.31%, *p* < 0.01). Meanwhile, tree age as another physiological factor also played a key role in leaf nitrogen isotope compositions (*R*<sup>2</sup> = 3.52%, *p* < 0.01). Moreover, looking at the relative contribution of each predictor in leaf nutrients, leaf nitrogen concentrations independently explained the largest percentage of variation in leaf δ15N (*R*<sup>2</sup> = 19.29%, *p* < 0.001).

**Figure 4.** Variation partitioning (%, *R*2) of physiological factors including leaf age, tree height and tree age; and leaf nutrients such as leaf nitrogen and phosphorus concentrations in accounting for leaf δ13C (**A**) and δ15N (**B**) values. The percentage values represent the proportion of the variance explained by each predictor in the model. \*\*\* *p* < 0.001, \*\* *p* < 0.01, \* *p* < 0.05.
