The way in which species richness varies with productivity has direct implications for the composition and functioning of ecological communities. However, the spatial scaling of this relationship remains unclear, as do the underlying mechanisms. To address this knowledge gap, we develop a quantitative, mechanistic theory of how the relationship changes with spatial grain (sample area) in a community of species competing for common resources.
Our theory was implemented using a neutral model community, in which species with the same demographic rates compete indirectly for common resources over a given spatial area. The model community loses species via stochastic extinction, but gains species via an influx of immigrants from a surrounding metacommunity. Productivity was measured as the rate of production of biomass per unit area. The neutrality assumption resulted in a parsimonious null model that is analytically tractable. We analyzed the model to determine how species richness varied with productivity at different spatial grains, as well as the underlying mechanisms. We also compared model predictions with trends found previously from a study of 25 forest plots, spanning temperate and tropical regions in five continents.
Results/Conclusions
At a given spatial grain, our model typically predicted a unimodal relationship between species richness and productivity. The increasing part of the relationship was driven by a sampling effect, whereby more species were sampled as the number of individuals increased with productivity. In contrast, the decreasing part was driven by a "dilution effect" on immigrants, whereby fewer species were able to establish via immigration due to the number of locally-produced recruits increasing with productivity. With increasing spatial grain, the area:perimeter ratio increased, resulting in relatively more locally-produced recruits. This increased the strength of the dilution effect, such that the peak of the unimodal relationship moved to the left. These trends from our model, derived from analytical solutions, were consistent with the observed spatial scaling of relationships from the forest plots.
Our results demonstrate that a unimodal relationship between species richness and productivity can arise from a neutral model, and does not necessarily imply strong niche effects. Furthermore, by showing that the peak of this relationship moved to the left with increasing spatial grain, our results suggest that species richness and productivity are more likely to be negatively correlated at larger spatial grains.