Biodiversity ensures ecosystem functioning and provisioning of ecosystem services, but it remains unclear how biodiversity–ecosystem multifunctionality relationships depend on the identity and number of functions considered. Here, we demonstrate that ecosystem multifunctionality, based on 82 indicator variables of ecosystem functions in a grassland biodiversity experiment, increases strongly with increasing biodiversity. Analysing subsets of functions showed that the effects of biodiversity on multifunctionality were stronger when more functions were included and that the strength of the biodiversity effects depended on the identity of the functions included. Limits to multifunctionality arose from negative correlations among functions and functions that were not correlated with biodiversity. Our findings underline that the management of ecosystems for the protection of biodiversity cannot be replaced by managing for particular ecosystem functions or services and emphasize the need for specific management to protect biodiversity. More plant species from the experimental pool of 60 species contributed to functioning when more functions were considered. An individual contribution to multifunctionality could be demonstrated for only a fraction of the species.
We thank E. Marquard, I. Kertscher, Y. Kreutziger, S. Rosenkranz and A. Sabais for providing additional data and T. Lewinsohn for comments on the paper. We thank the gardeners, technicians, student helpers and managers of the Jena Experiment for their work. The Deutsche Forschungsgemeinschaft (FOR 456 and FOR 1451) and Swiss National Science Foundation financed the Jena Experiment.
Electronic supplementary material
Supplementary material, Supplementary Tables, Supplementary Figures, Supplementary References
List of explanatory variables describing the plots of the Jena experiment. Given are plotcodes, spatial blocks, the sown species richness, the number of functional groups sown in the plot and information on the presence of individual functional groups in the plots
The measured values for the 82 ecosystem variables indicating functions in the 81 plots of the Jena experiment from the last year available as used in the main analyses
The measured values for the 54 ecosystem variables indicating functions in the 81 plots of the Jena experiment all measured in 2004 as used in a sensitivity analysis for the effect of including data on ecosystem functions measured in different years
Information for each ecosystem variable whether higher or lower values are considered indication of higher functioning
R code to calculate an index of multifunctionality using the introduced multivariate approach
R code to updating the turnover approach to multifunctionality to included more stringent criteria for considering an effect of an individual species informative. This allows to compensate for a high number of tests when many functions and many species are included in an analysis using the turnover approach. The code also includes analyses of simulated data to explore, based on permutation procedures, how large the potential bias in the results of the turnover approach are when using very large data sets
High-resolution version of the Supplementary Figure 2.3 showing loadings of all 82 ecosystem variables on principle components of the ordination calculated in the main analysis
High-resolution version of the Supplementary Figure 2.4 showing pairwise correlations between all 82 ecosystem variables included in the main analysis to indicate ecosystem functions and the calculated index of multifunctionality