2.5 Data analyses
All the data analyses were performed in R 4.1.3 (R Core Team, 2022), and all the figures were created using SigmaPlot 14.0. Prior to data analyzed, we natural log-transformed continuous variables to homogenize variance, except for temperature data because they may contain negative values. We used absolute values of P 50 for transformation because the original values were negative (Liu et al ., 2021). The differences in hydraulic and anatomical traits between roots and branches were tested using Welch’s t -test (Delacreet al ., 2017). Pearson correlation and linear regression were used to analyze the correlation between root and branch traits. One-way analysis of variance (ANOVA) was used to test the differences in meanP 50 root - branch for species across different biomes. Multiple regression and stepwise regression were used to select the appropriate factors for analyses (Hector et al ., 1999). Redundancy analysis (RDA; Braak, 1986) was used to determine the relative importance of environmental factors explaining hydraulic segmentation between roots and branches. Because of the multicollinearity among the environmental factors, we used variation and hierarchical partitioning (Chevan & Sutherland, 1991) and estimated unique and shared contributions among the factors. RDA analysis was performed using the rdacca.hp package (Lai et al ., 2022). The ‘rdacca.hp’ function was used in variation and hierarchical partitioning. The ‘permu.hp’ function was implemented significance testing for individual environmental factor contribution to explain variation of P 50 root - branch. For all analyses, the significant effects were inferred at P < 0.05.
Results