Landscape composition and configuration influence cereal aphid–parasitoid–hyperparasitoid interactions and biological control differentially across years

• Landscapes with more natural habitats enhanced both aphid density and parasitism.
• Small- vs. large-field landscape effects were non-significant or contradictory.
• Marginal vegetation type affects aphids and parasitoids differentially.
• Threshold range for effective biological control was 22–24% across landscape types.

The loss of landscape heterogeneity through agricultural intensification is known to affect aphid–parasitoid–hyperparasitoid interactions, with consequences for biological control. Various aspects of landscape heterogeneity (e.g. landscape composition and configuration) are expected to affect these interactions differentially, but there were few attempts to empirically compare the influence of separate landscape features on pest-parasitoid dynamics. To address these questions, we conducted three simultaneous studies in wheat fields in northern Serbia, to compare the effects of contrasting landscape contexts: (1) simple vs. more complex landscapes; (2) large- vs. small-field landscapes; (3) large-field areas with contrasting character of their marginal vegetation. We (1) found that aphid densities, parasitism rates and species richness of parasitoids and hyperparasitoids were higher in landscapes with more extensive and diversified non-crop habitats, positively affecting the biological control. We (2) did not find significant differences in aphid abundance and parasitism between large- and small-field landscapes, but we detected some contradictory patterns in aphid growth and parasitism increase; we relate both findings to certain region-specific landscape features of wider relevance. The character of marginal vegetation (3) had mixed effects on aphid–parasitoid interactions and dynamics, with respect to source of colonization. Parasitism rates above 22–24% were associated with population decline in the aphids, consistently across analyzed landscape contrasts. Other relationships were subject to significant interannual variability (over 2–4 years period), suggesting that effectiveness of landscape management for conservation biological control would also fluctuate year by year. Our findings show that a well-founded landscape-scale management for biological pest control in agriculture must be adjusted for differential aspects of landscape heterogeneity effects on pest–parasitoid interactions.