While there is considerable evidence indicating that increased biodiversity improves ecosystem functioning, most research has focused on small-scale studies which may not apply to conservation decisions at landscape scales. Theory on the scaling of biodiversity-ecosystem functioning (B-EF) relationships from small to larger scales has recently emerged, but few empirical studies of natural ecosystems have tested these predictions. Here, we investigate the scale dependence of the B-EF relationship using three metrics of tree-species diversity at different spatial scales (α-diversity, β-diversity, and γ-diversity) and aboveground tree biomass in a natural and environmentally-heterogenous forest ecosystem. We analyzed 14, 120 × 120 m (1.4 ha) landscapes of oak-hickory forest in the Missouri Ozarks, which each contain an environmental gradient of soil-nutrient and -moisture availability, light availability, and topographic conditions. Using a fixed grain size, we calculated the slope of the correlation between aboveground tree biomass and each of the three diversity metrics at 11 spatial extents within each landscape, ranging from 20 × 20 m to 120 × 120 m.
Results/Conclusions
We found that of the three metrics, β-diversity (variation in species composition among 10 × 10 m quadrats for a given spatial extent) was the best predictor of biomass at all scales. Moreover, both the slope and coefficient of determination (R2) of the relationship between β-diversity and tree biomass increased consistently with spatial extent, accounting for up to 50% of the variation in biomass at the landscape level. The slope of the relationship between γ-diversity (total species diversity at a given extent) and tree biomass also increased at larger spatial extents but showed little change at smaller spatial extents. Finally, α-diversity (average species diversity in 10 × 10 m quadrats in a given extent) was the weakest predictor of tree biomass, showing a consistently poor positive correlation with no explanatory power at all spatial extents. These results suggest that in natural, heterogenous landscapes, β-diversity is a stronger determinant of ecosystem function than either γ- or α-diversity, likely because it reflects important community assembly mechanisms that increase species sorting and turnover. Our study also provides evidence for scale dependency in the B-EF relationship in a natural ecosystem.