Fig. 4
Predicted partial least squares path modeling (PLS-PM) for the direct and indirect associations between SR and leaf stoichiometry. Ellipse represents the structural model, and box represents corresponding measurement models. In the structural model, the lines indicated paths, and the values adjacent to the lines denote the magnitude of the path coefficients calculated by PLS regression. R2 values are shown for all endogenous latent variables in the ellipse. Values in the measurement model represent the loadings between a latent variable and its indicators. The figure is showing the final models after model diagnosis processes. MAP, leaf N concentration, soil pH, SOC (soil organic carbon) and SN (soil N content) were removed because of low loadings (|loading| < 0.70). Pseudo Goodness-of-Fit (GoF) of the model is 0.570. All paths are significant except the paths from Climate to Soil, and Climate to Stoichiometry (p < 0.05)