Mapping differences in mammalian distributions and diversity using environmental DNA from rivers

Abstract

AimFinding more efficient ways to monitor, and estimate the diversity of, mammalian communities is a major step towards their management and conservation. Environmental DNA (eDNA) from river water has recently been shown to be a viable method for biomonitoring mammalian communities. Yet, most of the studies to date have focused on the potential for eDNA to detect individual species, with little focus on describing patterns of community diversity and structure. In this study, we focus on the sampling effort required to reliably map the diversity and distribution of semi-aquatic and terrestrial mammals and allow inferences of community structure surrounding rivers. LocationSoutheastern England MethodsWe used eDNA metabarcoding on water samples collected along two rivers and a beaver enclosure over two days, targeting terrestrial and semi-aquatic mammals. Mammalian community diversity and composition was assessed based on species richness and {beta}-diversity. Differences between river communities were calculated and partitioned into nestedness and turnover, and the sampling effort required to rapidly detect semi-aquatic and terrestrial species was evaluated based on species accumulation curves and occupancy modelling. ResultseDNA metabarcoding efficiently detected 25 wild mammal species from five orders in two days of sampling, representing the vast majority (82%) of the species expected in the area. The required sampling effort varied between orders, with common species (generally rodents, deer and lagomorph species) more readily detected, with carnivores detected less frequently. Measures of species richness differed between rivers (both overall and within each mammalian order) and patterns of {beta}-diversity revealed the importance of species replacement in sites within each river, against a pattern of species loss between the two rivers. Main conclusionseDNA metabarcoding demonstrated its capability to rapidly detect mammal species, allowing inferences of community composition that will better inform future sampling strategies for this Class. Importantly, this study highlights the potential use of eDNA data for investigating mammalian community dynamics over different spatial scales.