Scale Dependence of Bird Distributions in London

Understanding how spatial scale influences bird diversity and species-environment relationships is fundamental to effective conservation in increasingly urbanized landscapes, yet multiscale approaches remain underutilized in practice. Urban ecological studies often search for a single “scale of effect”—the spatial scale at which environmental factors best explain species occurrence. However, both theory and empirical evidence suggest that species respond to factors at multiple scales simultaneously, where single-scale perspectives may limit our understanding. To address this challenge, we examined bird diversity and occurrence over a range of scales across Greater London's urban-rural gradient through three complementary studies. Using crowdsourced (eBird) observations that included 279 species, we first applied scale variance analysis to identify how diversity variance changes across scales from 500m to 32km. We then modeled 65 species from the British Breeding Bird Survey using joint species distribution models with factors quantified across 5 scales (1,000m-16km) to assess how species-environment relationships change with scale (Part 2). Finally, we assessed factors across 10 scales (125m-64km) to determine which factors at which scales best explain occurrence for each species using a systematic selection model (Bayesian latent indicator scale selection). Scale variance analysis (Part 1) revealed that approximately two-thirds of bird diversity variance occurred at scales below 4,000m, with variance shifting from London's urban core to the rural periphery as scale increased. Joint species distribution modeling (Part 2) demonstrated that factors were highly scale-dependent, with at least one factor reversing direction across scales for 40% of the modeled species. Systematic factor selection (Part 3) revealed that environmental factors at local scales (125-500m) best explained occurrence for habitat specialists, while synanthropic species showed meaningful responses across all spatial scales, suggesting broader niche breadths. Collectively, the results illustrate that urban bird communities are structured hierarchically across scales, challenging the common practice of seeking a single best explanatory scale while supporting theories of hierarchical habitat selection. These findings provide both methodological advances for multiscale ecological analysis and practical guidance for designing urban conservation policies that better coordinate habitat management from individual properties to regional initiatives.