Resource-use efficiency drives overyielding via enhanced complementarity

Overyielding, the primary metric for assessing biodiversity effects on ecosystem functions, is often partitioned into "complementarity" and "selection" components, but this reveals nothing about the role of increased resource use, resource-use efficiency, or trait plasticity. We obtained multiple overyielding values by comparing productivity in a five-species mixture to expected values from its component monocultures at a) six levels of nitrogen addition (spanning 0-500 kg N ha(-1) year(-1)) and b) across four seasons. We also measured light, water, and nitrogen use, resource-use efficiency, and three functional traits-leaf nitrogen content, specific leaf area, and leaf area ratio-n mixtures and monocultures. We found strong evidence for non-transgressive overyielding. This was strongest in spring, with mixture productivity exceeding expected values by 20 kg dry matter ha(-1) day(-1). Peak overyielding was driven by enhanced complementarity, with the two non-N-2-fixing forb species far exceeding expected productivity in mixtures. Peak overyielding also coincided with higher water use in the mixture than for any monoculture, and enhanced mixture-resource-use efficiency. There was only weak evidence that trait plasticity influenced overyielding or resource use. Our findings suggest that when complementarity drives overyielding in grassland mixtures, and this is made possible both by increased water use and enhanced efficiency in water, nitrogen, and light use. Our results also suggest that mixtures offer a viable compromise between productivity, resource-use efficiency, and reduced environmental impacts (i.e., nitrate leaching) from intensive agriculture.