Résumé
Pest regulation in agroforestry systems (AFS) is beginning to be well-recognized, but the mechanisms implied in the interaction network between the environment, pests and predators in AFS are still not well-described. The aim of this work is to understand how plant diversity associated with AFS regulates the coffee berry borer (CBB) taking into account a tripartite interaction network: pest–predator–environment. It further seeks to understand how farmers' management practices can modify the regulating network. Using field data from coffee-based AFS and structural equation modeling, we assessed the effects of environmental conditions (% shade cover, tree area surface, coffee density) and farm management (conventional, integrated, organic) on (1) ant predatory groups, (2) the abundance and the damage from CBB and (3) their interactions. Percentage of shade cover was positively correlated to CBB initial infestation (through direct effect) and negatively through its effect on coffee phenology. A higher percentage of shade is also negatively related to damage intensity. Farmers’ practices significantly reduced the CBB population without considerable side effects on the ant predatory group, probably due to the high plant diversity within these farms. The abundance of the most diversified ant predatory group has a top-down effect on the peak of the CBB infestation rate. Our approach appears promising for a better understanding of the complex regulating network in coffee AFS and confirms the importance of an integrated management strategy to reduce CBB damage.
•The most diversified groups of predatory ants had a significant negative top down effect on CBB infestation.•Higher % of shade favours CBB infestation.•Regulation of CBB depends on the balance between the effects of local (in plot) and global (farm level) variables.•Farmers practices could be efficient to manage CBB and to reduce related damage.•Holistic approach is required to characterize more precisely regulating network in agroforestry systems.