Substrate-Dependent Variation in Environmental DNA Persistence and Degradation From a Small Mammal.

Environmental DNA (eDNA) analyses have been applied to terrestrial environments for over a decade, yet knowledge of the dynamics of eDNA deposition, movement and degradation remains limited. To address this, we conducted a controlled experiment over 1 year to assess the persistence of eDNA from an arboreal mammal across four substrates (water, sediment, sheltered nest-box surfaces, and exposed nest-box surfaces). Environmental DNA deposited by captive individuals was sampled repeatedly following removal of animals and analysed using both species-specific qPCR and metabarcoding. We modelled degradation using quantitative decay, detection probability, and conditional decay frameworks, and applied a random breakage model to estimate theoretical fragment persistence. Decay followed first-order exponential kinetics in water and on nest box surfaces but not in sediment, which exhibited high stochasticity. We observed a large difference in degradation rates between substrates, with exponential decay models revealing half-lives ranging from 7.3 days in water to 276 days on sheltered internal surfaces. Rapid decay in water restricted detections to approximately 77 days, while persistence models predicted that sheltered surfaces could retain detectable DNA for more than 4 years (> 1700 days), indicating the potential for prolonged accumulation of legacy DNA in protected environments. While species-specific qPCR offered higher sensitivity than metabarcoding, both methods had parallel decay patterns. Our findings indicate that substrate selection is a critical decision when designing an eDNA survey, as there is a trade-off between maximising detection probability and maintaining temporal resolution.