A spatiotemporal analysis of the effect of urbanization on birdsong

Rising population density and human activities have caused an increase in low-frequency ambient noise, which disrupts auditory communication systems, such as birdsong. In oscine songbirds, birdsong is a learned behavior that functions in territory defense, species recognition, and mate attraction. Increased noise levels in urban areas can reduce signal efficacy, which can have negative impacts on fitness. However, many species have persisted in urban settings by singing at less noisy times of day, singing at higher amplitudes, or changing aspects of their song to reduce masking by low-frequency ambient noise. Previous studies have often focused on modern populations of single species in rural and urban settings; here, we conduct a spatiotemporal analysis of the effects of urbanization on the birdsongs of nine species across ∼60 years. We use machine learning to assess distinguishability between songs from historic and modern timepoints. Additionally, we assess whether variation within modern songs is associated with urbanization, and whether song differences over time are associated with changes in urbanization levels. We find that while there are some song-feature differences between historic and modern songs, and between modern songs along a rural-urban gradient, these differences are not consistent across species. These results reinforce the notion that species respond differently to urbanization by altering different aspects of their songs. For the vocal-learning songbird species studied, historic and modern songs were not distinguishable using dimensionality reduction techniques and machine learning, and some modern songs differed in more urban settings. However, the non-vocal-learning (suboscine) species showed an overall increase in its lower song frequencies in modern songs, but showed no systematic differences in song features along an urbanization gradient, suggesting more dynamic changes in the vocal learners and population-level changes in the non-learners. This spatiotemporal study is the first to compile and analyze historic and modern songs from multiple species alongside urbanization data, bringing new insights to the effects of urbanization on birdsong.