Stronger effects of biodiversity loss and compositional change on forest biomass dynamics at larger spatial scales

Global environmental changes drive biodiversity loss and community compositional change. Yet whether and how both factors simultaneously impact biomass dynamics in natural ecosystems remains elusive, especially considering their multidimensional effects (e.g., taxonomic, functional, and phylogenetic) over spatial scales. To fill this knowledge gap, we generated an experimental spatial gradient using circular quadrats that vary in radius (2–30 m) in a subtropical forest on Dinghushan Mountain, China. Within each quadrat over 10 years, we calculated the changes in aboveground biomass (i.e., net Δbiomass), biodiversity (i.e., Δbiodiversity for richness, Shannon diversity, functional, phylogenetic), and community composition (i.e., β-diversity for taxonomic, functional, phylogenetic). Based on multi-model inference, we determined the most parsimonious relationships of Δbiomass as a function of Δbiodiversity and β-diversity and then quantified their standardized coefficients in response to the spatial gradient. Our results showed that Δbiomass, Δbiodiversity, and β-diversity decreased with quadrat size; the former at an accelerating rate and the latter at decelerating rates. While Δbiomass as a function of Δbiodiversity and β-diversity had low occurrences across the gradient, Δbiomass was strongly related to the change in functional dispersion (i.e., ΔFDis) and taxonomic β-diversity at larger spatial scales. Our results suggest scale-dependent influences of biodiversity loss and community compositional change on biomass dynamics in natural ecosystems. Further, our results highlight that multiple dimensions of biodiversity should be considered when predicting biomass dynamics at large spatial scales.