DNA metabarcoding and spatial modelling link diet diversification with distribution homogeneity in European bats

Abstract

Inferences of the interactions between species' ecological niches and spatial distribution have been historically based on simple metrics such as low-resolution dietary breadth and range size, which might have impeded the identification of meaningful links between niche features and spatial patterns. We analysed the relationship between dietary niche breadth and spatial distribution features of European bats, by combining continent-wide DNA metabarcoding of faecal samples with species distribution modelling. Our results show that while range size is not correlated with dietary features of bats, the homogeneity of the spatial distribution of species exhibits a strong correlation with dietary breadth. We also found that dietary breadth is correlated with bats' hunting flexibility. However, these two patterns only stand when the phylogenetic relations between prey are accounted for when measuring dietary breadth. Our results suggest that the capacity to exploit different prey types enables species to thrive in more distinct environments and therefore exhibit more homogeneous distributions within their ranges.

Fig. 1. Dietary diversity statistics of the analysed bat species.

a Radial phylogenetic tree of prey detected using the Zeale primers and their occurrence patterns in each of the studied bats. A higher resolution image (Supplementary Fig. 1) and the homologous figure built from the Epp data (Supplementary Fig. 2) are available in the Supplementary Information. b Overall and predator species-specific representation of the arthropod taxonomic orders. The incidence-based figure is available as Supplementary Fig. 4. c Dietary niche breadth measures accounting for richness (dR), richness + evenness (dRE) and richness + evenness + regularity (dRER). The error bars (±SE) of dRER indicate the dispersion of the dietary niche breadths yielded when using different prey phylogenetic trees (N = 50) sampled from the Bayesian MCMC. d One-dimensional species ordination plots ranking species according to the dietary niche breadth based on different metrics. Levin’s index is also included for being the most common metric employed in the literature.

Fig. 2. Relationship of species’ dietary niche breadth measures with spatial features.

a–g Species distribution models (SDMs) of the seven bat species generated using an ensemble approach. The bright area represents the distribution area according to the IUCN Red List. h Significant linear relationship between the recognised range size measured from IUCN maps (x-axis) and the potential range estimated from the SDMs (y-axis). i Non-significant linear relationship between distribution homogeneity and distribution range. j Non-significant linear relationship between dietary niche breadth accounting for the three components of diversity (dRER: richness + evenness + regularity) and recognised range size. k Non-significant linear relationship between dietary niche breadth accounting only for dietary richness (dR) and distribution homogeneity. l Significant linear relationship between dietary niche breadth accounting for the three components of diversity (dRER) and distribution homogeneity. Dots indicate mean values per bat species. Note that error bars (±SE) in charts (j) and (l) indicate the dispersion of the different dietary niche breadth values yielded from the 50 iterations run with different prey phylogenetic trees to account for phylogenetic uncertainty.

Fig. 3. Relationships of dietary, spatial and behavioural traits of species.

a Dietary alpha diversity of species shows no linear correlation with dietary niche breadth. b Dietary niche breadth exhibits a significant linear correlation with prey turnover. c Distribution homogeneity shows a borderline linear trend with prey turnover. d Dietary niche breadth exhibits a significant linear correlation with hunting plasticity.