Non-structural carbohydrates (NSC) play important roles in adapting to environments in plants. Despite extensive research on the seasonal dynamics and species differences of NSC, the relative contributions of season and species to NSC is not well understood. We measured the concentration of starch, soluble sugar, NSC, and the soluble sugar:starch ratio in leaves, twigs, trunks and roots of twenty dominant species for dry and wet season in monsoon broad-leaved evergreen forest, respectively. The majority of concentration of NSC and starch in the roots, and the leaves contained the highest concentration of soluble sugar. A seasonal variation in starch and NSC concentrations higher in the dry season. Conversely, the wet season samples had higher concentration of soluble sugar and the sugar:starch ratio. Significant differences exist for starch, soluble sugar and NSC concentrations and the sugar:starch ratio across species. Most species had higher starch and NSC concentrations in the dry season and higher soluble sugar concentration and the sugar:starch ratio in wet season. Repeated variance analysis showed that starch and NSC concentrations were strongly affected by season although the effect of seasons, species, and the interaction of the two on the starch, soluble sugar, and NSC concentrations were significant.

It is essential to understand how the loss of biodiversity impacts both ecosystem function (EF) and multifunctionality (EMF). Previous studies have mostly focused on predicting how species richness (SR) impacts EMF, while the effect of functional diversity (FD) on EMF remains unclear. Specifically, we know little about the primary functional drivers impacting EMF compared with SR. Therefore, we analysed 8 ecosystem functions within 58 natural secondary forest plots to investigate the effect of FD on both individual EF and EMF. Our results suggest that SR and FD had very significant positive effects on plant phosphorus, soil available phosphorus, and soil total nitrogen. FD explained significantly more variations in these functional responses than SR for individual ecosystem functioning. We also used a multiple threshold approach to test the effect of SR and FD on EMF. We found that FD and SR were positively related to EMF regardless of whether low-level function or high-level function was desired, but FD had a larger effect than SR. Based on the averaging approach, OLS regression, multivariate linear regression model and random forest analysis, we found that SR and FD were both drivers of EMF but that FD had a stronger effect and could explain more variation. As such, we conclude that FD drives ecosystem multifunctionality more than SR.