Abstract
Introduction Environmental DNA (eDNA) metabarcoding of water is increasingly being used to monitor coastal biodiversity shifts. However, we have limited knowledge of whether samples collected during discreet temporal periods depict holistic ecosystem changes over longer time spans. Methods Here, we show how eDNA community structure varies across repeated sampling events at different temporal scales ranging from years to months to days at an Arctic coastal site, Churchill (Canada), using metabarcoding analyses of water eDNA samples with four universal primer pairs (two primers in COI and two in the 18S rRNA). Results Daily variations were highly dynamic and less structured, likely due to the stochastic nature of estuarine ecosystems, but there was a clear annual consistency in eDNA communities with a high proportion of shared taxa between years. However, monthly sampling was the most efficient for capturing holistic biodiversity. Discussion We provide recommendations for optimal eDNA metabarcoding sampling design based on our observations. The study underscores the importance of understanding biological and physical factors altering eDNA detection to improve the efficiency of detecting and interpreting long-term eDNA changes.