Linking biodiversity-production relationships to community assembly: A mechanistic modelling approach

Krishna Anujan and Shahid Naeem, Ecology, Evolution and Environmental Biology, Columbia University, New York, NY, Robert Bagchi, Ecology & Evolutionary Biology, University of Connecticut, Storrs, CT, Sebastian Heilpern, Natural Resources and the Environment, Cornell University, Ithaca, NY

Background/Question/Methods
The relationship between biodiversity and primary production in a natural ecosystem is often positive but highly variable in shape and strength. The success of conservation and carbon capture policies like REDD+ depend on improving the accuracy of these predictions and building stronger mechanistic links between biodiversity and biomass production. Processes that generate diversity gradients like dispersal, environmental heterogeneity and interspecific competition are good predictors of species composition and traits in a community but are often not explicitly considered in predictions of biomass production. In this context, we ask what is the contribution of trait-based community assembly processes to the diversity-production relationship?
We use probabilistic mechanistic models of tree community assembly to model the colonization of a set of sites from a mainland through three sequential processes (i) seed dispersal, (ii) survival in the abiotic environment and (iii) biotic competition. We model the probability of a species to exist in a site as a set of functions of both species’ traits and site properties. We then derive the diversity-production relationship and test the sensitivity of this relationship to the strengths of the processes and the underlying gradients.
Results/Conclusions
Preliminary results show that with a regional richness of 500 (mean abundance per unit area = 150), a species richness (S) gradient is established across sites (0 to ~400 species). Richness has an increasing and saturating effect on primary production (logS coefficient = 0.6, p<0.01), a pattern broadly consistent with observations. Increasing the density of dispersers on all sites uniformly, we find the shape of the curve changes from saturating to linear to exponentially increasing (logS coefficient=1.9, p<0.01) for the same set of site conditions. Further, we find that the gradient of environmental heterogeneity and the extent of dispersal limitation affect the strength of the relationship, with steeper gradients leading to higher exponents. But within the range of our parameters, these gradients have little effect on the shape of the relationship. In conclusion, our preliminary results show that significant differences in the shape and strength of diversity-production relationships could result from differences in community assembly processes predictable from site properties. Our models can be applied to set realistic expectations for regional-scale patterns on field.