Fig. 3: Variations in leaf N, P, and N:P explained by species identity and environmental factors in random forest (RF), linear regression (LM), and linear mixed effect models (LMM). | Nature Communications

Fig. 3: Variations in leaf N, P, and N:P explained by species identity and environmental factors in random forest (RF), linear regression (LM), and linear mixed effect models (LMM).

From: Environmental versus phylogenetic controls on leaf nitrogen and phosphorous concentrations in vascular plants

Fig. 3

a, d, g Proportion of variation in the data aggregated to the site-level explained by least squares regression models (LM) and RF. b, e, h Proportion of within-species variations explained by LM and RF models. c, f, i Proportion of variation in the full data explained by linear mixed-effect models (LMM) and RF models. The brown bars are determined as the intraclass correlation coefficient (ICC) from LMMs and as the cross-validation R2 from RF models that contain only species identity, family, and genus information as predictors, but no environmental variables. The green bars on top of the brown bars represent the difference to the full models, where environmental variables and species identity were used as predictors. Source data are provided as a Source Data file.

Back to article page