Many studies have revealed a positive effect of biodiversity on temporal stability thereby supporting the “insurance hypothesis.” Such effects are typically explained by interspecific niche differentiation, leading to differential responses to the environment and thus compensatory dynamics among species, buffering the impact of environmental changes. However, how to identify and quantify the drivers of the stabilizing insurance effect remains an open question. Loreau & de Mazancourt (Am. Nat. 2008) identified three such drivers: intra- and interspecific density dependence, environmental forcing and demographic stochasticity.
Our aim is to quantify the contributions of each of these three drivers in a long-term field experiment. We used species specific aboveground biomass from 2002–2012 to investigate temporal patterns of community stability. We constructed hierarchical models, including the drivers of community dynamics, to test the effects of synchrony (as covariance of the communities), dominance and response to climatic fluctuations on the stability of plant communities across a diversity gradient. We then used model comparisons to select those models best explaining our data.
Results/Conclusions
From preliminary results it appears that environmental factors (e.g. temperature and precipitation aggregated with PCA) do not drive temporal stability of our grassland ecosystem. Instead, community structure (e.g. evenness of the community), the summed covariances and the diversity of the community together with their interaction terms, explain most of the variability in temporal stability of aboveground productivity. Overall, species do fluctuate asynchronously, but these fluctuations are primarily driven by the different responses of communities dominated by few species compared to homogeneously composed communities with higher evenness. Moreover, a prevalence of negative summed covariances would be expected when compensatory dynamics are taking place, but we find evidence for an opposite trend, with the majority of covariances being positive. This suggests that other factors (like trait differences or environmental factors that have not yet been considered) are most likely the drivers of community stability rather than compensatory dynamics.