The urgent need for understanding the consequences of ongoing species losses has generated many experiments testing the effects of biodiversity on selected ecosystem processes. Most such studies have been conducted in artificially assembled, early successional plant communities, and have addressed responses in ecosystem process rates and community composition as if they were largely independent. Therefore, how functional responses to biodiversity loss are influenced by compensatory dynamics from remnant species and invasion from exotic species remains largely unexplored. In a large-scale removal experiment, we evaluated the short-term effects of losing different plant functional groups on grassland primary productivity and invasibility in the Flooding Pampa of Argentina. We established eight blocks, each comprising 11 3x6 m experimental plots. Plots were randomly assigned to different treatments including the selective removal of five different functional groups (native forbs, native winter grasses, native summer grasses, exotic forbs, or exotic grasses), four levels of random biomass removal (5, 20, 40, or 80%), and two intact grassland controls. During three growing seasons, beginning one year after the initial removals, we periodically harvested above- and below-ground plant biomass from two 20x100 cm quadrats in each plot, and estimated aerial net primary productivity (ANPP) and total root mass. 

Results/Conclusions

The removal of each functional group significantly decreased ANPP during the first two growing seasons. The magnitude of the removal effect on productivity was positively related to the initial biomass of the different plant groups in the system. After four years, remnant species compensated for the loss of minor functional groups. In contrast, ANPP did not recover to control levels where native summer grasses or all exotic (mostly winter) grasses had been removed. Removing native summer grasses also decreased belowground community biomass. Overall, effects of selective functional group removals were similar to those produced by randomly removing equivalent amounts of plant biomass. However, removal of native summer grasses led to a four-fold increase in productivity of exotic winter forbs, an effect not found under random biomass disturbance. We conclude that the loss of dominant functional groups constrained energy inputs in accordance with a ‘mass-ratio’ model of ecosystem functioning. In contrast, selective perturbation of a dominant native plant group may increase invasibility by exotic species unable to functionally compensate for the loss of the native residents. This could have potential consequences on soil water and nutrient retention in the ecosystem.