Semi‐natural habitat, but not aphid amount or continuity, predicts lady beetle abundance across agricultural landscapes

Abstract The amount of semi‐natural habitat surrounding farm fields is a common but inconsistent predictor of natural enemy populations and predation services. Standard land cover metrics may not accurately capture the actual availability of limiting resources for natural enemies and can miss important dynamics across space and time. Theory from animal movement and landscape ecology predicts that regions with more, spatio‐temporally continuous resources (i.e. food and shelter) should have larger predator populations and enhanced biological control. To test these predictions empirically, we designed a study measuring aphids, lady beetles and predation services in agricultural landscapes in Wisconsin, USA. In two study years, we sampled lady beetles and aphids in 336 crop fields (corn, soybean, alfalfa and small grains) and adjacent semi‐natural habitat patches (grasslands and woodlands) across 24 1.5 km buffer landscapes at four to seven time points each, and in 1 year, we assessed predation rates with sentinel egg cards. We used aphid counts to model habitat‐specific aphid phenologies, from which we calculated landscape indices of prey amount and continuity. These indices, along with semi‐natural habitat area, were used to predict lady beetle abundance. While there were strong differences in the abundance and timing of aphids by habitat, semi‐natural habitat amount was still a better predictor of lady beetle counts and sentinel egg predation than either aphid amount or continuity indices in these landscapes. Synthesis and applications . Our findings confirm the robust relationship between lady beetles and semi‐natural habitat in agricultural landscapes, and highlight the complexities of measuring fine‐scale resource heterogeneity in real landscapes. Retaining or adding woodland and grassland patches in agricultural landscapes is likely to support larger lady beetle populations and enhance predation in crop fields. Our results suggest that these habitats may be more important for shelter than prey continuity, though this mechanism warrants further investigation. Future work should continue to refine experimental methods for the successful integration of landscape ecology and animal behaviour to support conservation goals.