Combining environmental niche models, multi-grain analyses, and species traits identifies pervasive effects of land use on butterfly biodiversity across Italy

Understanding how species respond to human activities is paramount to ecology and conservation science, one outstanding question being how large-scale patterns in land use affect biodiversity. To facilitate answering this question, we propose a novel analytical framework that combines environmental niche models, multi-grain analyses, and species traits. We illustrate the framework capitalizing on the most extensive dataset compiled to date for the butterflies of Italy (106,514 observations for 288 species), assessing how agriculture and urbanization have affected biodiversity of these taxa from landscape to regional scales (3–48 km grains) across the country while accounting for its steep climatic gradients. Multiple lines of evidence suggest pervasive and scale-dependent effects of land use on butterflies in Italy. While land use explained patterns in species richness primarily at grains ≤12 km, idiosyncratic responses in species highlighted “winners” and “losers” across human-dominated regions. Detrimental effects of agriculture and urbanization emerged from landscape (3-km grain) to regional (48-km grain) scales, disproportionally affecting small butterflies and butterflies with a short flight curve. Human activities have therefore reorganized the biogeography of Italian butterflies, filtering out species with poor dispersal capacity and narrow niche breadth not only from local assemblages, but also from regional species pools. These results suggest that global conservation efforts neglecting large-scale patterns in land use risk falling short of their goals, even for taxa typically assumed to persist in small natural areas (e.g., invertebrates). Our study also confirms that consideration of spatial scales will be crucial to implementing effective conservation actions in the Post-2020 Global Biodiversity Framework. In this context, applications of the proposed analytical framework have broad potential to identify which mechanisms underlie biodiversity change at different spatial scales.