Forest productivity (increment of above-ground biomass) is determined by biodiversity but also by stand structure attributes. However, the relative strengths of these drivers in determining productivity remain controversial in subtropical forests. In this study, we analyzed a tree growth data from 500 plots with in a 20 ha mature subtropical forest in eastern China. We used spatial simultaneous autoregressive error models to examine the effects of diversity variables (species richness, evenness, and composition), stand structural attributes (stand density, tree size range and diversity), environmental factors (topography and soil), and initial above-ground biomass (AGB) on productivity. We also applied structural equation models to quantify the relative importance of diversity, stand structure, environmental factors, and initial AGB in determining forest productivity. Our results showed that stand structure together with diversity and initial AGB governed forest productivity. Tree size diversity (DBH Shannon’s diversity index) had the largest positive effect on forest productivity. These results provide new evidence that structural explanatory variables have greater contributions to productivity for mature subtropical forests, strongly supporting the niche complementarity hypothesis. Our work highlights the importance of tree size diversity in promoting high forest productivity, and suggests that regulating and conserving complexity of forest stand structure should be among the most important goals in subtropical forest management.

Leaf economic spectrum (LES) traits (e.g., leaf mass area and leaf nutrient concentrations) are effective indicators of an acquisitive or conservative resource use strategy for plants. The increased atmospheric deposition of nitrogen (N) and phosphorus (P) alters soil nutrient availability, thereby affecting plant LES traits. However, how the LES traits and their trade-offs affect the resource-use strategies of understory plants under long-term N and P additions is still unclear. Based on a fertilization plot including four treatments (control (CK), N addition (+N, 100 kg N hm⁻¹ yr⁻¹), P addition (+P, 50 kg P hm⁻¹ yr⁻¹), and combined N and P additions (+NP, 100 kg N hm⁻¹ yr⁻¹ + 50 kg P hm⁻¹ yr⁻¹)) conducted over 12 years in a subtropical broad-leaved evergreen forest, this study addresses the differential response of four LES traits (leaf mass per area (LMA), leaf N concentration (Nmass), leaf P concentration (Pmass) and leaf net photosynthesis per unit mass (Amass)) to fertilization in five dominant understory plants (Camellia fraternal, Eurya muricata, Eurya rubiginosa, Rhododendron ovatum, and Symplocos sumuntia) to test whether trade-offs between plant traits closely related to light resources play an important role in influencing plant resource-use strategies. The results show that, compared to the CK treatment, the LMA was significantly increased by 12.5% to 12.8% under +N treatment, and the Nmass was significantly increased by 25.9% and 23.6% under +N and +NP treatments, while Pmass and Amass were significantly increased by about 23% and 15~50%, respectively, after P addition. There was a highly significant negative correlation between the response of LMA and Pmass, irrespective of the addition of N and P alone or together. The increase in LMA under +N treatment made the resource-use strategy of the understory plants more conservative. Meanwhile, the understory plants tended to rapidly acquire resources by decreasing LMA while increasing Pmass under +P and +NP treatments. Our results suggest that, under long-term N and P additions, understory plants with limited light availability change their resource-use strategies mainly through the trade-off between leaf LMA and Pmass, which should be considered to capture the long-term adaptive strategies of understory plants against a background of intense atmospheric N and P deposition.