Using traits to integrate non-additive effects of species mixtures across ecosystem processes

Whether and how biodiversity affects ecosystem functioning has long been hotly debated in ecological research and conservation. Important in this debate is how interactions between species in a community lead to non-additive effects (i.e. effects that deviate from predictions based on the effects of each single species) on ecosystem processes. Such non-additivity has been widely reported for individual processes, for example productivity, decomposition, fire, herbivory or pollination. However, species in a community are simultaneously involved in multiple ecosystem processes. We therefore propose a trait-based conceptual approach to connect non-additive effects based on species interactions across different specific ecosystem processes and illustrate its potential. The approach involves plotting the direction and strength of non-additivity due to species interaction effects for any given ecosystem process against the values of relevant predictive traits for all possible pairs of species considered in a community. Synthesis: We show how to compare the non-additivity patterns for different ecosystem processes using similar ‘currency’ and how to link these to the main effects of the same species on these ecosystem processes. This way the species' effects on higher-level ecosystem functioning (e.g. carbon cycling), in present and future environmental scenarios, can be better quantified. The conceptual framework requires empirical testing and incorporation of relevant environmental factors.