The influence of ecological complexity on the interactions between key ecosystem services

Chatura Vaidya, Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI and John Vandermeer, School for Environment and Sustainability, University of Michigan, Ann Arbor, MI

Background/Question/Methods
Pollination and pest-control are vital ecosystem services mostly studied in isolation. Their availability depends on management practices that may favor one ecosystem service over another in order to increase yields. In this study, we investigated how pest-control and pollination interact with each other in coffee agroecosystems, and how they are mediated by soil fertility due to nitrogen-fixing shade trees which are managed in the system. The ant, Azteca sericeasur, is a known biological control agent in coffee agroecosystems whose impact on coffee pests is mediated through a mutualistic association with the green coffee scale (GCS) insects that feed on coffee plants. The presence of A. sericeasur on coffee presents the necessary conditions for a potential trade-off between pest-control and pollination, as its activity on plants may potentially deter pollinators. Furthermore, nitrogen-fixing trees in the system can potentially modify this interaction by changing the strength of the A. sericeasur-GCS mutualism with varying quality of honeydew due to supplemental nitrogen. Using a full factorial design, we selected coffee bushes within 2m of shade trees that were either nitrogen-fixing trees (Inga micheliana) or non-nitrogen fixing trees (n=32 each) with half containing nests of A. sericeasur.
Results/Conclusions
Coffee plants closer to Inga trees had higher variability in A. sericeasur activity, largely due to ants tending an alternative species of scale insect which feeds in the canopy of Inga trees. This observed shift in the foraging of A. sericeasur due to the nitrogen-fixing trees, did not significantly impact infestation of the pest of coffee, the coffee berry borer, but sites with A. sericeasur had lower infestation of coffee berry borer by 0.5%. Furthermore, we found that pollinators preferred to visit flowers on plants without A. sericeasur and whenever they did visit plants with A. sericeasur, their visit duration was significantly less, suggesting that A. sericeasur may be deterring pollinators to some extent. A. sericeasur was also more likely to forage for nectar in flowers in the absence of GCS, and although pollinators seem to prefer ant-free plants, the effect of ants was not enough to impact fruit-set and fruit weight. Overall, we found some supporting evidence for the potential for trade-offs in ecosystem services in coffee agroecosystems, and highlight the role of number of complex indirect interactions determining ecosystem services.