Yann Clough

Control of crop pests by shifting host plant availability and natural enemy activity at landscape scales has great potential to enhance the sustainability of agriculture. However, mainstreaming natural pest control requires improved understanding of how its benefits can be realized across a variety of agroecologies. Empirical studies suggest significant but highly variable responses of natural pest control to land-use change. Current ecological models are either too specific to provide insight across agroecosystems, or too generic to guide management with actionable prediction. We propose instead to harness the joint benefits of empirical, theoretical and methodological knowledge for natural pest control modelling across systems by combining trait-mediated understanding from correlative studies with the explicit representation of causal relationships achieved by mechanistic modelling. To link these frameworks, we adapt the concept of archetypes, or context-specific generalizations, from sustainability science. Similar responses of natural pest control to land-use gradients across cases that share key attributes, such as functional traits of focal organisms, indicate general processes that drive system behaviour in a context-specific manner. Based on such observations of natural pest control, a systematic definition of archetypes can provide the basis for mechanistic models of intermediate generality that cover all major agroecosystems worldwide. Example archetype models based on general ecological theory and knowledge of American and African agroecosystems explain responses of natural pest control to changes in landscape composition and configuration across Europe. Analysing global datasets on drivers and components of natural pest control and mechanistic models of the resulting archetypes can resolve inconsistent responses to land-use change, and improve prediction of natural pest control potential, as well as synergies and trade-offs with other ecosystem services.