Humans are propelling drastic environmental changes leading to biodiversity loss at regional and global scales. Locally, higher species richness often improves the ability of ecosystems to function. Thus, continuing biodiversity loss may have devastating consequences for ecosystem functioning. These potential consequences of biodiversity loss underpin biodiversity-ecosystem functioning research in local-scale experiments. In naturally assembled systems however, loss in species richness at the local scale does not appear to be the dominant pattern. Further, in naturally assembled systems, the effects of biodiversity change on ecosystem functioning depend on processes at multiple scales. Understanding the effect of biodiversity change on ecosystem functioning requires scaling theory that bridges from these local experiments to large-scale naturally assembled communities. Here, we use macroecological patterns like the Species-Area-Curve to present a null hypothesis for how biodiversity ecosystem functioning relationships should scale across naturally assembled landscapes. Further, we compare these predictions to our expectations in experiments and validate these predictions with data from a nutrient poor grassland and a seasonally try tropical forest.
Results/Conclusions
We find that the relationship between species richness and ecosystem functioning must change with increasing sampling extent because species richness increases non-linearly with increasing sampling extent. Whether this change increases the importance of species richness for ecosystem functioning depends on how ecosystem functioning is calculated. If ecosystem functioning is calculated as a per area measure as it is in biodiversity experiments, then the slope of the biodiversity-ecosystem functioning relationship should decrease with increasing sampling extent. If ecosystem functioning is calculated as a summed measure as it is in many theory studies of scaling biodiversity-ecosystem functioning relationships in naturally assembled systems, then the slope of the biodiversity-ecosystem functioning relationship should increase with increasing sampling extent. Further, we find that the underlying macroecological patterns in experiments may be very different from those in naturally assembled systems and thus our expectations should differ.