FIGURE 6.

Effects of forest structural attributes on light heterogeneity. (a) Height of inflection point (HIP) of relative light intensity versus HIP of cumulative basal area, (b) light attenuation rate (i.e. the slope of light extinction) versus the slope of the cumulative crown area, (c) relative light intensity at 1 m versus total crown length per plot and (d) horizontal light heterogeneity at 1 m (expressed as the standard deviation in relative light intensity across 16 subplots at 1 m) versus horizontal heterogeneity of crown area (expressed as the coefficient of variation (%) in the crown area). In (a, b and c), grey dots represent the value of each of the 16 subplots, and black dots represent the average value across the 16 subplots per plot. The results of a regression line (black line) and coefficient of determination (r: Pearson correlation coefficient and τ: Kendall's Tau) are shown. Pearson correlation coefficient was obtained using linear mixed model setting plots as a random factor. Kendall's Tau was obtained using the nonparametric Mann–Kendall trend procedure, and the slope and intercept of its regression line were obtained using Sen's protocol. HIP of relative light intensity and the cumulative basal area is the height of inflection point, and light attenuation rate and slope of cumulative crown area are the slopes at the inflection point, as obtained by fitted sigmoidal curves per plot. Regression lines are (a) y = 1.07x − 2.03; (b) y = 0.96x + 0.19; (c) y = −0.19x + 4.56; (d) y = 0.70x − 0.69