Quantifying insecticide exposure across an agricultural landscape gradient using barn swallows as a wildlife...

Agricultural intensification and rising insecticide use are known threats to biodiversity in agroecosystems. Using a series of agricultural and landscape metrics, we tested the hypothesis that barn swallows (Hirundo rustica), as an avian wildlife indicator species, are at greater risk of insecticide exposure in more agriculturally intensive landscapes. We used spatial datasets for insecticide use density, landscape simplification, and relative cover of row crops and aquatic habitats (wetlands and other water bodies) at multiple spatial scales (500 m-20 km) as proxies of agro-intensity. In 2022 and 2023, we captured and blood sampled 173 barn swallows from 54 farms across a large agricultural gradient in southern Saskatchewan, Canada, and collected invertebrate sweep samples from three habitat types (cropland, grass or pastureland, and wetland margins) on 37 farms for chemical analysis to determine detection prevalence and sum concentrations of seven systemic insecticides (imidacloprid, clothianidin, thiamethoxam, acetamiprid, thiacloprid, flonicamid, and dinotefuran). The neonicotinoids thiamethoxam, clothianidin, and imidacloprid were detected in 54%, 47%, and 27% of the 173 barn swallow plasma samples and 98% of the 120 invertebrate samples. Insecticide concentration and detection in swallow plasma declined over the season but showed no clear or strong relationship with the measured metrics of agricultural intensity. However, the detection or concentration of insecticides in invertebrate samples positively correlated with most agricultural metrics. These results are consistent with the foraging behavior of barn swallows, whose movement across several kilometers likely integrates variation in prey contamination across farmland. The lack of association of agricultural intensity with barn swallow plasma neonicotinoid levels and positive relationship with invertebrates indicates that either rapid metabolism of these compounds in birds prevents accurate exposure characterization using blood, or land-use intensity maps are insufficiently precise at the relevant spatial scales for a wide-ranging species, and/or lower detection frequency of insecticides precludes characterization of fine-scale insecticide risk. We advocate for finer scale spatial data on agricultural intensity by crop type and pesticide use and recommend broader invertebrate sampling to more accurately reflect pesticide exposure risk for wildlife.