Do biodiversity indices behave as expected from traits of constituent species in simulated scenarios?

It is believed that diversity of plant communities has a positive effect on their productivity. The benefits of diversity are described by “biodiversity indices”, comparing yield of mixtures with yields of monocultures of constituent species. These indices are supposed to capture also the main mechanisms leading to increased yield. We have constructed a spatially explicit individual based model, simulating even-aged stand development, and compared the behaviour of selected biodiversity indices (overyielding, selectivity and complementarity) with expectations based on life history traits of constituent species. The results are based on comparisons of two species mixtures with corresponding monocultures. We designed three scenarios of changes in the two species life history differentiation, and compared the behaviour of the indices with expectation based on it. In the first scenario, selectivity was driven by increasing size inequality of the two species, mostly in accordance with expectations. The second scenario presents increasing shade tolerance of the smaller species that increased complementarity, again mostly as expected. In the last scenario, shortening of length of stress tolerance of the weaker species surprisingly increased values of the biodiversity indices. For each setting, we varied sowing density and spatial pattern of the constituent species. The behaviour of the indices was influenced by both factors, but the effect of density was more pronounced. In particular, at high sowing densities, the most important interactions happened in the very early stages of mixture development, and the behaviour of the indices was often counter-intuitive.

► Biodiversity indices should reflect mechanisms of productivity response to diversity.
► We used individual based models to simulate plant monocultures and mixtures.
► Sowing density and spatial pattern qualitatively influenced the values of indices.
► At lower densities, the indices reflected well the trait differentiation of species.
► At very high densities, the behaviour of indices was counter-intuitive.