Sampling Design and Considerations for Environmental DNA Collection in Aquatic Systems

Thursday, September 12, 2013: 2:00 PM
Fulton (Statehouse Convention Center)
Timothy Strakosh , Green Bay Fish and Wildlife Conservation Office, US Fish and Wildlife Service, New Franken, WI
John Sweka , USFWS, Northeast Fishery Center, Lamar, PA
Chris Olds , USFWS, Alpena Fish and Wildlife Conservation Office, Alpena, MI
Stephen Hensler , USFWS, Alpena Fish and Wildlife Conservation Office, Waterford, MI
William Chadderton , The Nature Conservancy, Notre Dame
Christopher Jerde , Environmental Change Initiative, Department of Biological Sciences, University of Notre Dame, Notre Dame, IN
The application of environmental DNA in aquatic systems for detecting rare species is rapidly advancing both in technology and use. However, few studies have focused on the sampling design for eDNA collection. Using conventional study designs (i.e., random) for species with low detection probabilities is often cost and logistically prohibitive.  Targeted sampling approaches (i.e., focusing on likely eDNA accumulation areas) can reduce effort and increase the probability of detection.  However, several factors need to be considered to maximize the detection probability of eDNA, including target species behavior and eDNA dispersal and persistence.  Knowing species life history and behavioral characteristics can identify times and locations the target species may congregate and release higher amounts of DNA into the environment (e.g., spawning). Dispersal of eDNA will be dependent on hydrologic characteristics of the area, dictating where eDNA will likely accumulate in the surface film (e.g., behind velocity breaks and downwind areas). Environmental factors (e.g., temperature, UV, microbial activity, etc.) will influence eDNA degradation rates and reduce detectability. Targeted sampling designs combine a priori knowledge while adapting for in situ conditions.  These can be merged with conventional designs to increase spatial coverage, while maximizing detection probabilities and minimizing the number of samples collected.