Local scale crop compositional heterogeneity suppresses the abundance of a major lepidopteran pest of cruciferous vegetables

Previous studies have recommended preserving semi-natural habitats as a strategy to promote natural enemies and reduce the abundance of agricultural pests. Such non-crop habitats, however, may increase pest abundance, causing spillover from non-crop to crop fields. A potentially more economical and attractive solution for farmers might arise if crop fields were themselves designed to aid in pest control. As part of such a strategy, we examined whether increasing crop compositional heterogeneity (i.e. the number of crop types and their evenness) and crop configurational heterogeneity (i.e. field margin length) could reduce the abundance of Pieris canidia, a major agricultural pest butterfly in several Asian countries. Adult females of P. canidia lay their eggs on cruciferous vegetables (Brassicaceae), and the emerging larvae consume the leaf tissue, causing crop damage. We surveyed adult P. canidia across 52 farmlands in south China when their abundance was at a peak in spring (April to May). Our results showed that the percentage of cruciferous crops (cabbage, oilseed rape, and pak choi) at the 100 m radius local scale was strongly positively associated with P. canidia abundance. However, crop compositional heterogeneity, also at the local scale, significantly reduced the abundance of P. canidia. Field margin type, categorized by the spatial coexistence of the most dominant cover types (sugarcane, corn, vegetables, including crucifers, and weedy vegetation), was also an important explanatory variable, with weedy patches, usually consisting of fallow cropfields adjacent to the field margins, having the highest pest abundance. These results suggest that to control pest attack by P. canidia on cruciferous vegetables, increasing crop compositional heterogeneity could be a more effective strategy than increasing configurational heterogeneity or the amount of non-crop habitat. However, research measuring crop damage by larval butterflies is necessary to confirm this hypothesis.