Fungal communities influenced the relationship between tree diversity, traits, and productivity
The SEM that included fungal community composition (Fig. 1B; Table S3) explained 45% of the variability in tree productivity compared to 30% for the null SEM (Fisher’s C = 39.882, P < 0.001 on 6 df; Fig. 1D). Tree functional diversity (std. coef. = 0.13, P = 0.011) and functional identity (std. coef. = -0.13, P = 0.012) influenced fungal community NMDS axis 1, which was positively related to productivity (std. coef. = 0.24, P = 0.005). Tree species richness (std. coef. = -0.21, P = 0.005) and tree functional identity (std. coef. = -0.65, P < 0.001) negatively affected fungal community NMDS axis 2, which was negatively related to productivity (std. coef. = -0.33, P < 0.001). Tree functional identity (std. coef. = -0.14, P = 0.059) had a marginal negative effect on fungal community NMDS axis 3, which was positively related to productivity (std. coef. = 0.19, P = 0.002). The magnitude of the direct effects of fungal community composition on productivity were similar to the direct effects of species richness (std. coef. = 0.18, P = 0.028), functional diversity (std. coef. = 0.20, P = 0.018), and functional identity (std. coef. = -0.35, P < 0.001) on productivity. These relationships between fungal community composition and tree diversity resulted in positive, indirect effects for tree species richness, functional diversity, and functional identity on productivity.