Abstract
Environmental DNA (eDNA) surveys are increasingly used for biodiversity monitoring, providing unprecedented resolution and sensitivity for understanding biological community composition. These surveys are especially important for monitoring non-indigenous species, for which early detection and mitigation actions are of critical importance. However, eDNA studies focussing on the effects that different environmental sample types have on community composition remain rare. Here we used eDNA metabarcoding of 18S rRNA and COI genes to unravel community composition and to compare sediment and water samples. We first detected markedly different communities and a greater number of distinct operational taxonomic in sediment compared to water. We then compared the obtained datasets with previously published biodiversity surveys and found excellent concordance among different survey techniques. Finally, eDNA metabarcoding detected many non-indigenous species including several novel species introductions. We conclude that careful consideration on the type of environmental sample is needed when conducting eDNA surveys, especially for assessments of community change.