Spatial mismatch between wild bee diversity hotspots and protected areas.

Wild bees are critical for multiple ecosystem functions but are currently threatened. Understanding the determinants of the spatial distribution of wild bee diversity is a major research gap for their conservation. Here, we model wild bee α and ß taxonomic and functional diversity in Switzerland to (i) uncover countrywide diversity patterns and determine the extent to which they provide complementary information, (ii) assess the importance of the different drivers structuring wild bee diversity, (iii) identify hotspots of wild bee diversity, and (iv) determine the overlap between diversity hotspots and the network of protected areas. We use site-level occurrence and trait data from 547 wild bee species across 3343 plots and calculate community attributes, including taxonomic diversity metrics, community mean trait values, and functional diversity metrics. We model their distribution using predictors describing gradients of climate, resource availability (vegetation), and anthropogenic influence (i.e. land-use types and beekeeping intensity). Wild bee diversity changes along gradients of climate and resource availability, with high-elevation areas having lower functional and taxonomic α-diversity and xeric areas harbouring more diverse bee communities. Functional and taxonomic ß-diversities diverge from this pattern, with high elevations hosting unique species and trait combinations. The proportion of diversity hotspots included in protected areas depends on the biodiversity facet, but most diversity hotspots occur in unprotected land. Climate and resource availability gradients drive spatial patterns of wild bee diversity, resulting in lower overall diversity at higher elevations, but simultaneously greater taxonomic and functional uniqueness. This spatial mismatch among distinct biodiversity facets and the existing degree of overlap with protected areas challenges wild bee conservation, especially in the face of global change, and calls for better integrating unprotected land. The application of spatial predictive models represent a valuable tool to aid the future development of protected areas and achieve wild bee conservation goals. This article is protected by copyright. All rights reserved.