How, what, and where you sample environmental DNA affects diversity estimates and species detection

Environmental DNA (eDNA) is a complex mixture of DNA, varying in particle sizes and distributed patchily in aquatic systems. Optimizing eDNA sampling is crucial for maximizing species detection, particularly in high-risk scenarios like invasive species management. This study compares two eDNA capture methods (grab sample vs. tow net) at the opposite ends of the spectrum for volume and particle size to ascertain what most influences 1) the detection of invasive species (Dreissenid mussels and Burmese pythons) and 2) total diversity monitoring of metazoan, fungi, and protist with COI marker and plant communities with an ITS2 marker. Sampling was conducted across a wide geography and diverse aquatic environments in Minnesota and Florida, USA and Switzerland. The tow net samples were consistently higher in eDNA yield compared to grab samples; however, they exhibited equal or lower alpha diversity of ZOTUs (operational taxonomic units), challenging an assumption that there is a relationship between eDNA yield and diversity. We found biodiversity patterns were significantly influenced by the capture method, especially for COI communitymetazoan diversity in all three regions. The two capture methods measured different beta diversity of COImetazoan communities across all three regionlocations, highlighting the confounding impact of sample volume and filtration pore size on the diversity of eDNA captured. Interestingly, the beta diversity of plant eDNA was less impacted by the capture method compared to themetazoan beta diversity of COI communities. We found no clear difference in detection for the two invasive species targets with respect to the eDNA capture method. These results underscore the need for pilot studies before conducting biodiversity inventory and monitoring, and a need for a greater understanding of not just how much, but also what eDNA is captured depending on method choice, considering both spatial and particle size heterogeneity. For biodiversity inventories, further research is needed to understand what causes differences in detection, and what are the optimal eDNA sampling strategies.