From bird communities to trophic interactions: successive ecological filters decouple potential and realized insectivory in urban environments.

1. Urbanization is widely associated with functional homogenization of ecological communities, including declines in insectivorous bird species and concomitant restructuring of arthropod assemblages. While these changes are often assumed to alter ecosystem functioning through trophic interactions, it remains unclear whether bird insectivory in cities can be inferred from community composition alone, or whether it also depends on which species actively forage, on natural prey availability and on the arthropod guilds exploited. 2. We hypothesized that urbanization (i) restructures both insectivorous bird communities and arthropod prey assemblages, thereby (ii) increasing divergence between trait-based potential and realized insectivory, while (iii) prey use is shaped by natural prey availability and active prey selection. 3. We used 25 plots distributed along a controlled urbanization gradient in Montreal, Canada. We combined acoustic monitoring of insectivorous bird communities, arthropod surveys, and camera-monitored cafeteria-like experiments deployed on 73 trees to quantify trait-based potential insectivory, realized insectivory through direct foraging interactions, natural prey availability and prey selection among three arthropod guilds (lepidopteran larvae, spiders and ants). 4. Urbanization reduced bird functional diversity while also decreasing the availability of high-quality prey, particularly lepidopteran larvae and spiders. Realized insectivory was consistently decoupled from community-based expectations, with actively foraging assemblages systematically less insect-specialized than predicted by broader community compositions (potential insectivory). This mismatch was non-random and only weakly related to urban and vegetation gradients. Prey use was largely independent of natural prey availability and instead reflected consistent preferences for larval prey over other guilds, with some indication that the functional composition of bird foraging assemblages further shaped which prey were exploited. 5. Our results show that realized trophic interactions emerge through successive ecological filters rather than directly from community composition: environmental conditions restructure bird communities and prey availability, subsequent foraging filters determine which species actively participate, and prey selection governs which guilds are consumed. Predicting ecosystem functioning in cities therefore requires considering not only biodiversity patterns but also the ecological processes linking community composition to realized trophic interactions.