The landscape context of natural enemy interactions in a South Korean agroecosystem: Implications for the effectiveness of biological pest control

Background/Question/Methods

TERRECO is an interdisciplinary project aiming to optimize the long-term provision of ecosystem services in the Haean agricultural catchment and the Soyang Lake watershed, in northern South Korea. In particular, it investigates the role of landscape structure for the abundance and distribution of natural enemies in crop fields, in relation to their effectiveness for biological pest control. Landscape structure is a key determinant of biodiversity in agricultural landscapes, and may thus represent an essential interface for local optimization of biological control. To date, however, it is unclear whether landscape effects on higher trophic levels induce differences in the degree of pest suppression, damage avoidance, or crop yield. In addition, interactions between enemies, for instance linked with omnivory of certain guilds, might also vary according to the landscape, with consequences for the effectiveness of control.

In this study, the effects of landscape context on natural enemies and on their interactions are examined through the lens of their consequences for pest density, herbivory, and crop yield. In 18 fields distributed throughout spatially distinct landscapes, seven treatments excluding various combinations of natural enemy guilds were installed on organically-managed cabbage plants (Brassica oleracea var. capitata).

Results/Conclusions

Results show that all treatments accessible to natural enemies had lower leaf damage, lower pest densities and higher final biomass than exclusion controls. In addition, the strength of pest control effects varied depending on the landscape, particularly in the case of flying insect enemies (syrphid flies, parasitoid and predatory wasps), who reduced herbivory rates by 1.1% in simple landscapes and by 30.6% in complex landscapes. However, damage reduction in complex landscapes was lower in the presence of birds than in treatments excluding them, thus pointing to negative interactions between guilds. These results imply 1) that quantification of the biological control potential requires consideration of ecosystem service-oriented, crop-level variables, as it is not deducible from patterns of enemy abundance only; 2) that a broader landscape perspective on these variables is necessary in order to quantify this service in real-world ecosystems. This may provide key guidelines both for landscape-scale management of biological control, and for higher predictability of multitrophic interactions.