Reply to Mokany et al: Comprehensive measures of biodiversity are critical to investigations of ecosystem multifunctionality.

Mokany et al. (1) reiterate an important consideration about the interpretation of results in our PNAS report (2). The authors suggest that the use of a different metric of β diversity might reveal a larger effect of β diversity on ecosystem multifunctionality. We agree, and reiterate that our reported effects of β diversity on ecosystem multifunctionality are likely conservative for a number of reasons. We also emphasize the importance of our findings despite these conservative results, and encourage future studies to explore creative methods for capturing more biodiversity effects. Mokany et al. (1) specifically hypothesize that spatially explicit metrics of β diversity would have stronger effects on ecosystem processes and lead to a larger insurance effect than we observed. We agree that this is likely. To create replicate experimental landscapes and take advantage of the strengths of a highly controlled, long-term experiment, we used two nonspatial β diversity metrics. As we discussed, two of the eight functions we measured were likely influenced by spatial orientation (insect richness and insect abundance). Thus, the use of nonspatial β diversity metrics likely underestimated β diversity’s true impact. In addition to spatial considerations, we also hypothesized that future studies will reveal stronger effects of each scale of biological diversity (α, β, and γ diversity) if they incorporate more realistic community structure and larger gradients in community composition (2). Given the constraints imposed by the original experimental design of this study, it is remarkable that two simple (but very different) nonspatial metrics of β diversity produced measurable and consistent impacts on ecosystem multifunctionality. In particular, higher β diversity decreased the variability in ecosystem multifunctionality, and the effects of both β and γ diversity only became apparent when multiple functions were analyzed simultaneously. In other words, even under the simplified constraints of this experimental design and these biodiversity metrics, it is clear that β diversity contributes to a multifunctional insurance effect, and that multiple functions must be considered to understand the total effects of biodiversity at multiple scales. We hope that future studies develop new methods to incorporate even more comprehensive measures of biodiversity to more fully assess its effects on ecosystem multifunctionality.