Fig. 4: Effect of climatic factors on the trait associations of forest communities (Principal component (PC) scores: PC1 in upper panels and PC2 in lower panels).
From: Mapping multi-dimensional variability in water stress strategies across temperate forests
a and d Result of the relative importance analysis of the individual climate variable for the trait associations along PC1 or PC2. A total of 12,413 forest communities are selected and the number of bootstrap samples is 1000 in the relative importance analysis. Violin plots show the probability density of the bootstrapped importance values. The boxplots indicate the interquartile range (the first quartile (Q1) and the third quartile (Q3)), which includes the median (50th percentile) of the bootstrapped estimates of the relative importance values for the five climatic variables. The whiskers represent the minimum and maximum values within 1.5 times the interquartile range from Q1 and Q3, respectively. b and e The linear regressions with the aridity index, as hypothesised in the introduction. A total of 12,433 forest communities are used in the analysis. c and f The regressions with the most important climatic variable for the PC1 (mean summer temperature: Summer_T) and PC2 (mean annual temperature: MAT) separately. The regression tests are two-sided and the regression lines and shading are presented as the means and standard errors, respectively. A total of 12,434 forest communities are used in the analysis. The shading in panels b, c, e and f follows the same colour legends as in Fig. 3 for PC1 and PC2, respectively.
