Stand structural attributes exert stronger and scale-dependent control on forest biomass than tree diversity across typical forest ecosystems in China

Understanding how biodiversity-ecosystem functioning (BEF) relationships scale spatially and temporally remains critical under global change. Here, using continuous monitoring databases across temperate, subtropical and tropical natural forests in China from 2004 to 2020, we quantified multiscale tree diversity-biomass dynamics. Linear mixed-effect and structural equation models were used to further disentangle the direct and indirect pathways through which tree diversity influences forest biomass at different scales. The results revealed significant positive linear relationships between tree species richness and stand biomass at 1200 and 400 m 2 scales ( P < 0.05), whereas no significant association was detected at 100 m 2 scale ( P > 0.05). Notably, these demonstrated relationships exhibited consistent temporal stability across all spatial scales throughout the study period. Importantly, the direct effect of tree diversity on biomass intensified with increasing spatial scale, while indirect effects mediated through stand structural attributes (CV_DBH and tree density) became proportionally stronger at finer scales. Further analyses showed that stand structural attributes emerged as the strongest predictor of biomass variation across all scales, surpassing both diversity and climate effects. Furthermore, spatial variation of climate factors (mean annual temperature and mean annual precipitation) mainly affected stand biomass through the indirect effects on tree species diversity and stand structural attributes. Overall, multiscale analyses revealed stand structural attributes dominates biomass prediction, with climate acting indirectly. Scaling biodiversity-structure strategies can enhance forest resilience under global change. Future work should integrate cross-scale mechanisms into climate-smart afforestation for sustainable carbon sequestration.