Community composition coupled with habitat fragmentation drives acoustic divergence in bird assemblages

Rapid expansion of human activities has altered abiotic and biotic environments and reshaped the sensory systems of animal species. Auditory perception, a key sensory component of soniferous species, is essential for signal detection, species recognition and group coordination. Birds mitigate acoustic masking in fragmented habitats by actively modulating the spectral and temporal features of their songs. However, it remains challenging to determine whether these modulations are primarily driven by biotic factors (e.g. species interactions within varying community compositions) or by abiotic factors (e.g. island attributes). We surveyed bird communities on forested islands in the Thousand Island Lake region, China, using passive acoustic monitoring. We applied sound frequency-based analyses to examine the relationship between island attributes (area and isolation), acoustic assemblage composition (species richness, morphological and phylogenetic relatedness) and sound frequency modulation of birds on 12 islands. Our findings indicated that species competition within acoustic space led to various strategies of frequency modulation to avoid acoustic overlap. With increasing frequency overlap, birds exhibited greater variations in peak frequency and frequency range, reflecting decoupled modulation in which vocal adjustments occurred either upward or downward, depending on context. By disentangling the effects of community composition from island attributes, we found that acoustic overlap was intensified on remote and small islands, particularly among species with large body size or close phylogenetic relationships, driving acoustic niche partitioning. These findings highlight the importance of biotic interactions within animal communities in driving avian vocal production shifts, emphasizing the necessity of jointly considering community composition and gradients of abiotic factors when examining sensory adjustments.