Landscape diversity enhances parasitism of cotton bollworm (Helicoverpa armigera) eggs by Trichogramma chilonis in cotton

•Influence of land use on parasitism of Helicoverpa armigera eggs was investigated.•Trichogramma chilonis was the only parasitoid found in cotton.•Landscape diversity enhances the parasitism.•A high proportion of urban habitats and water increases the biocontrol services.

We studied the effects of landscape composition and diversity on parasitism of Helicoverpa armigera eggs (Noctuidae: Lepidoptera) by Trichogramma spp. (Hymenoptera: Trichogrammatidae) in cotton fields in northern China. Sentinel eggs of H. armigera were exposed to parasitoids for a 48 h time periods in experiments conducted in 2012–2013, resulting in parasitism rates ranging from 0 to 38.8%. Trichogramma chilonis was the sole species of parasitoid wasp found in parasitized H. armigera eggs during the study period. Using open-access satellite imagery from Google earth, we classified the proportion of the landscape surrounding each field at radii of 0.5, 1.0, 1.5 and 2.0 km. Landscape variables were first analyzed using principal component analysis (PCA), and then general linear models containing all combinations of principal components 1–3 (PC1–3) and landscape diversity measured using Simpson’s Diversity Index (D) were compared using an adjusted Akaike’s Information Criterion (AICc). We found that the best fit model to explain detected variation in H. armigera egg parasitism varied by spatial scale. Parasitism rates were best predicted at 2.0 km indicating that parasitoids are interacting with the surrounding habitats at a relatively large spatial scale. At this scale the PC2 + PC3 model had the lowest AICc value. Parasitism rates were significantly positively correlated with PC3, indicating that high proportion of urban areas and water within the landscape at a 2.0 km scale leads to higher parasitism rates in cotton fields, whereas landscapes dominated by cotton and maize led to low rates of parasitism within a given cotton field. A non-significant negative correlation with PC2 was found. The landscape diversity (D) model had the lowest AICc values at spatial scales of 0.5–1 km and D + PC3 was the best fit model at 1.5 km. Cotton fields within diverse landscapes with a high proportion of urban habitats and water resulted greater parasitism of H. armegera than those embedded in agriculturally dominated landscapes. This study provides a primary understanding of the relationship between landscape variables and ecosystem services for insect pest management in Chinese small farming systems.