Detecting Invasive Round Goby in Wadeable Streams: A Comparison of Gear Types

Abstract

The invasion of round goby Neogobius melanostomus into tributaries of the Laurentian Great Lakes poses concerns for the conservation of native fish communities and the functioning of stream ecosystems. In streams, where standard methods for detection are lacking, development of sensitive sampling methods to identify round goby-invaded sites is a prerequisite for the implementation of effective management strategies. We sampled 34 stream sites that spanned a range of goby densities with two active gears (electrofishing and seining) and one passive gear (minnow traps) to compare gear efficiency at detecting invasive round goby. Naive occupancy, defined as the number of sites where goby were detected with each gear divided by the total number of sites, was 0.79 (27/34 sites) for seining, 0.68 (23/34) for electrofishing, and 0.50 (17/34) for traps. The probability of detecting a single round goby in a single pass (determined with single-season, constant-probability models) was 0.75 ± 0.065 (mean ± SE) for minnow traps, 0.69 ± 0.056 for seining, and 0.47 ± 0.075 for electrofishing. Mean ± SE catch per unit effort (CPUE; round goby/min) was 0.716 + 0.158 for seining, 0.137 ± 0.043 for electrofishing, and 0.078 ± 0.022 for minnow traps, seining being significantly more efficient than minnow traps and electrofishing. The CPUE did not differ between electrofishing and traps. Mean goby size did not differ among gear types, but a size bias was detected when data from a related study were included in the analysis, seining capturing smaller round goby. Based on the sensitivity and efficiency of seining, we conclude that when stream conditions allow, this gear is a practical means of detecting round goby and determining their abundance in streams.

Received August 11, 2011; accepted January 4, 2012

ACKNOWLEDGMENTS

We thank Brockton Feltman and Josh Martin for field help, and Alan Dextrase for advice about the use of PRESENCE and the models used. We also thank George Gamboa for the donation of a field vehicle and Cathy Starnes, Jan Bills, Rita Perris, and Sheryl Hugger for administrative and logistical support. Funding for this project was provided by Fisheries and Oceans Canada and Oakland University Provost's Graduate Student Research Award; JHGN was also funded through a teaching stipend in the Department of Biological Sciences at Oakland University.