Grassland biodiversity experiments have shown that random species loss at local scale reduces ecosystem productivity. Would the same apply in forest biodiversity experiments? We answered this question with a highly replicated experiment in subtropical forest in south-east China. Using a species pool of 42 tree species we created communities with 1, 2, 4, 8 and 16 species in such a way that every species occurred at all diversity levels the same number of times. Small plots contained 400, large plots 1600 trees. Large plots were divided into quarters to add a shrub-diversity treatment planted with 0, 2, 4 and 8 species. We measured tree basal area, height and volume (basal area x height) of the 16 central trees in each parcel of 400 trees over years 4-7 of the experiment and summed these measurements for each species and for the whole community to stand-level values.
Results/Conclusions
Stand-level basal area and volume increased linearly with the logarithm of tree species richness and this effect became stronger with time. Additive partitioning of the increasing net biodiversity effects over time showed that complementarity effects were always positive whereas selection effects were initially negative (species with low monoculture performance dominating mixtures) but increased over time to zero or positive values. The number of tree species with positive productivity responses to increasing tree species richness at population level increased over time. Plot size did not influence biodiversity-productivity relationships. Low shrub diversity (2 species) reduced tree productivity but high shrub diversity had a neutral or even positive effect on tree productivity. Our results demonstrate that the positive and over time increasing biodiversity effects on ecosystem productivity reported previously from grassland biodiversity experiments also apply to forests biodiversity experiments. Furthermore, biodiversity effects did not depend on plots size (again consistent with the absence of plot-size effects reported from grassland biodiversity experiments) but extended from tree to shrub species richness. Converting the experimentally found biodiversity-induced productivity gains into accumulated standing biomass suggests that planting diverse rather than mono-specific forests could greatly increase their contribution to carbon uptake and reducing global warming.