Results
Overall, 23 species were enumerated in the mixed species stands studied, with an average of 5 species per plot, and range value of 1 to 12 species. AGB was in average 15.5 Mg across all plots, and varied from 9.3 to 23.1 Mg per plot. The relationship between taxonomic diversity and AGB varied with specific metric. Among the four taxonomic diversity metrics, only species richness influenced positively AGB (Figure 1), and the effects were shown by significantly higher values of AGB in plot with more species (r2=0.16; p<0.05; Figure 1). As for Shannon diversity, Pielou evenness and Simpson index, they showed no clear pattern with AGB (r2 < 0.04; p>0.05; Figure 1). As such, species richness was the only taxonomic diversity metric predictor of AGB.
From the Multi-Model Inference results, single trait-functional diversity and functional dominance metrics were not retained in the final selected optimal model (Table 1; Figure 2). Only FEve and structural diversity metrics were retained as predictors for AGB (Table 1), and their effects were shown by increasing plot-level AGB with higher FEve and structural diversity, especially CV DBH and CV Npb (Figure 2). Although being retained in the final model, CV Ht showed positive but nonsignificant effect on AGB (Figure 2). These results were confirmed when examining the relative importance variable for all the diversity metrics included in the model (Figure 3), which indicated that the most important predictors of AGB were FEve, CV DBH and CV Npb.