Biodiversity affects several ecosystem functions and services, including carbon cycling and retention. While the efficiency of carbon capture by plant assemblages often correlates with species diversity, mutualistic interactions in plant-animal networks could further influence this important ecosystem function. The diversity mutualistic interactions and the connectance of the resulting multitrophic networks may be associated with higher numbers of generalist interaction partners, a redundancy that could provide ecosystem resilience in the face of disturbance and prevent sharp drops in primary productivity. Here, we ask whether animal-plant mutualisms modulate the biodiversity-productivity relationships observed for plants. Specifically, we use spatially-explicit structured equation models to test for the effect of the diversity of plant-animal interactions on regional estimates of NPP, while controlling for the underlying effect of the abiotic environment. To do so, we combine field observations of fruit-frugivore interactions in a biodiversity hotspot – i.e., the Brazilian Atlantic Forest, to spatial estimates of species potential distributions, providing the first biome-level surface of interaction likelihood and community complexity.
Results/Conclusions
Including biotic predictors in NPP models improved model fit (in terms of AIC and BIC). Plant diversity emerged as the strongest driver of the distribution of NPP, but animal-plant interactions magnified the positive effect of the biotic component on ecosystem productivity. Specifically, trophic network connectance was found to scale-up to affect the geography of ecosystem functioning, perhaps via dispersal-mediated spatial insurance effects. Our research supports the view that multiple dimensions of community complexity—including mutualistic interactions—contribute to ecosystem productivity in a biodiversity hotspot. Our work provides the first line of evidence that plant-animal interactions may shape the spatial distribution of zoogeochemical cycles via mechanisms other than consumption.