SRA

Changes in land use following habitat loss, particularly through agricultural intensification, is a primary driver in the decline of pollinators. This decline and accompanying increased isolation between populations can reduce genetic diversity, which may reduce adaptive potential and unmask deleterious genetic load. Maintenance of functional connectivity (i.e. gene flow) between populations in changing landscapes is necessary to ensure long-term persistence. However, how specific land uses impact genetic diversity and functional connectivity in pollinators remains poorly understood. Here, we assess to which extent grassland butterfly populations are functionally isolated by different land use types, and what consequences isolation has had for genetic diversity and connectivity using whole genome and land cover data. To understand species-specific responses to land use, we assess three butterfly species that vary in degree of grassland habitat specialization and mobility - Polyommatus icarus, Plebejus argus and Cyaniris semiargus - and sample these from 6-11 populations across a ~25,000 km2 area in southern Sweden. We find that the three species vary in functional connectivity, with the generalist P. icarus forming a largely panmictic population, P. argus having slightly elevated differentiation between sampled populations, and grassland specialist C. semiargus having largely isolated populations. Genetic diversity is positively related to grassland extent in the surrounding landscape in all species, however, arable and forest extent has opposing relationships with genetic diversity in P. icarus and C. semiargus. Low genetic diversity and high population differentiation relative to the other two species is coupled with higher rates of inbreeding in C. semiargus, suggesting more urgent conservation attention is needed to restore necessary functional connectivity in this species. In conclusion, maintenance of butterfly genetic diversity relies upon preservation of heterogeneous landscapes incorporating both semi-natural grasslands and forests, and genetic data can reveal threats that may go unnoticed with species abundance data alone.