| Abstract: | To develop a better understanding of how biodiversity loss and productivity are related, we need to consider ecologically realistic rarity (i.e. reduced evenness and increased dominance) and extinction (i.e. reduced richness) scenarios. Furthermore, we need to identify and better understand the factors that influence species and community yielding behaviors because the general conditions for overyielding are the same as those for coexistence. We established experimental tallgrass prairie plots in Iowa to determine how two ecologically realistic rarity–extinction scenarios influenced aboveground net primary productivity (ANPP) and disassembly. Equal‐mass seedlings of six tallgrass prairie species were transplanted into field plots to establish realistic declining species evenness (high, medium, low) and richness (4, 1) treatments. Across declining evenness treatments, the relative abundance of the ubiquitous tall species Andropogon gerardii increased, the relative abundance of the tall species Salvia azurea was constant, and the relative abundance of two short (dissimilar height scenario) or two tall species (tall scenario) decreased. Monocultures of Andropogon represented a continuation of this trend until there was complete dominance by Andropogon and extinction of all other species. Our treatments also allowed us to test if variation in plant height contributes to the complementarity effect. Niche partitioning in plant height was not positively related to complementarity. The effects of declining species evenness and richness on the diversity–productivity relationship were different for these two ecologically realistic rarity–extinction scenarios. Specifically, as diversity declined across treatments, ANPP and the selection effects decreased in tall communities, but not in dissimilar communities. Additionally, differences between these two scenarios revealed that decreased species yielding behavior is associated with two tallgrass prairie extinction risk factors, rarity and short height. The differences between these scenarios demonstrate the importance of incorporating the known patterns of diversity declines into future studies. |
