Doubting dung: eDNA reveals high rates of misidentification in diverse European ungulate communities

Abstract

Pellet counts are widely used to monitor ungulates but rely on the assumption that pellets of different species are correctly identified in the field. Recent studies question this assumption using DNA barcoding techniques to check field identification rates. For Europe, which is undergoing a rapid shift towards more diverse ungulate assemblages, such an assessment is still missing. Using DNA barcoding on 3889 fecal samples from nine ungulate species in four European countries, we found average field misidentification rates varied from 0.6% for horse (Equus ferus) to 41.1% for roe deer (Capreolus capreolus). Most identification errors occurred between similar-sized species from the same taxonomic family. For a subset of samples from Sweden, we looked at the effect of dung morphometry, observer experience, and season on species identification success. Dung morphometry clearly distinguished moose (Alces alces) but not red (Cervus elaphus), roe, and fallow deer (Dama dama). Experienced observers performed better than novices for red and fallow deer although still making significant identification errors (26% and 17% incorrectly identified). Identification success was higher during spring and winter (x¯$$ \overline{x} $$= 86%) than summer and autumn (x¯$$ \overline{x} $$= 74%). We question pellet counts as an accurate monitoring tool where similar-sized species coexist and monitoring relates to the whole community. For this increasingly common situation across Europe, DNA testing or camera traps may be a better alternative. Pellet counts remain useful where only few species with clearly different dung morphology coexist (e.g., moose and roe deer) or when focused on species with distinctive dung morphology (e.g., moose).