Filtered, yet not homogenised: Urbanisation does not lead to cross-city convergence in Mexican urban bird communities.

Urbanisation acts as a non-random filter of biodiversity, often resulting in less diverse communities dominated by a subset of species with particular traits. These patterns have led to the expectation that urban environments may promote biotic homogenisation, yielding taxonomically or functionally similar communities across cities, relative to their surrounding natural ecosystems. Here, we evaluated whether the non-random filtering of biodiversity by urbanisation results in the homogenisation of bird communities across three environmental conditions (i.e. natural ecosystems, urban greenspaces, urban sites) in three biogeographically distinct Mexican cities. Assuming an urban-driven filtering process (i.e. an overall decline in native species richness from natural ecosystems into the urban matrix, with possible additions of exotic species), we formulated scenario-based expectations to disentangle urban filtering from biotic homogenisation, predicting whether filtering results in locally distinct urban communities, or additionally, convergent filtering across cities drives cross-city homogenisation. We quantified taxonomic and functional richness (alpha diversity) and compositional dissimilarity (beta diversity) across natural ecosystems, urban greenspaces and urban sites to evaluate whether only filtering occurred or if such process has additionally led to context-dependent convergence in community composition across cities, which could indicate homogenisation. We analysed taxonomic and functional approaches using incidence- and abundance-based metrics to assess compositional variation within and among cities. We recorded 131 bird species, of which only a handful were shared across urban conditions in all three cities. Bird communities showed strong taxonomic and functional differentiation among cities, with greater similarity within the studied environmental conditions per city region than across cities, indicating that urbanisation filters species locally, with no clear evidence of cross-city convergence in community composition. Our findings indicate that urbanisation filters species locally, simplifying communities, yet avian assemblages remain more similar within the three environmental conditions per city region than across the three surveyed cities in Mexico. Notably, widespread and abundant urban-associated species contribute to increased similarity within urban assemblages but do not result in cross-city convergence. This challenges the notion that urbanisation causes homogenisation (at least across the three studied Mexican cities and their surrounding natural ecosystems, which belong to highly distinct biogeographic regions) and highlights the predominant influence of regional species pools and local environmental contexts in shaping urban avifaunas.