Species richness is not a good predictor for above-ground biomass in a warm temperate deciduous broadleaf forest

Background: Biomass is the result of long-term production and metabolism in forest ecosystems and is an important indicator of the carbon storage capacity of forests. Although there is increasing empirical evidence supporting the positive impact of biodiversity on forest productivity and biomass, there is still uncertainty about the relative importance of tree diversity in determining carbon storage compared to other factors such as environmental conditions, functional characteristics and stand structure, especially in complex forest ecosystems. Methods: In this study, based on dataset from a 25-ha forest dynamics monitoring plot, we investigated the effects of tree diversity, environmental variables, functional traits and stand structural attributes on above-ground biomass (AGB). Spearman correlation coefficients were used to analyze the correlations between AGB and the variables. The relative importance of these factors in influencing AGB variation was assessed using a random forest model. Structural equation model was used to determine the direct or indirect effects of each factor on AGB. Results: The results showed a negative, though not significant, correlation between species richness and AGB. There was a significant positive correlation between leaf dry matter content and leaf tissue density, implying that more leaf photosynthetic products were utilized for dry matter accumulation. The variation in AGB was mainly explained by the maximum diameter at breast height and the coefficients of variation of diameter at breast height, suggesting that large diameter individuals contribute disproportionately to AGB. In addition, AGB was also influenced by topographic factors (i.e., altitude and slope), while there was no significant correlation with soil variables. Conclusions: This study reflects the response of AGB to different influencing factors. Our study emphasizes that stand structure attributes may be more suitable as predictors of forest AGB than species richness.