In silico experiments with natural enemy response to landscape heterogeneity and resulting crop pest loads

Field studies and recent meta-analyses have found conflicting effects of habitat diversity on biological control in agricultural landscapes. To explore the reasons behind these variable responses, a detailed mechanistic model of pest/predator dynamics has been developed to enable in silico experimentation. The model structure allows for varying landscape factors such as coverage and distribution of natural habitats, varying resources provided by both natural habitats and crops, as well as varying predator traits such as overwintering densities, dispersal distances and predator search area. Overall, crop pest loads are highly variable across simulation scenarios, mimicking the conflicting results from field research. Simulation results suggest that while natural enemy response to landscape structure is moderated to some extent by dispersal behavior, variation in resource availability to early colonizers is pivotal to understanding the highly variable responses of natural enemies to landscape heterogeneity. In silico experimentation that informs, and is informed by, field research has the potential to yield predictions regarding when conservation biological control efforts are likely to be most successful.