You are here:
Early Detection Monitoring for Invasive Fish: St. Louis River (SLR) Pilot Study
Citation:
PETERSON, G. S., J. HOFFMAN, A. S. TREBITZ, C. W. WEST, AND J. KELLY. Early Detection Monitoring for Invasive Fish: St. Louis River (SLR) Pilot Study. Presented at Minnesota Invasive Species Conference 2008, Duluth, MN, October 26 - 28, 2008.
Impact/Purpose:
To recognize that early detection of aquatic invasive species is necessary to develop and implement timely management responses.
Description:
Early detection of aquatic invasive species is necessary to develop and implement timely management responses. Predicting species introductions, however, is difficult and resources are typically limited. Therefore, monitoring strategies should be designed to effectively and efficiently address a suite of potential invaders. Strategies must consider life histories, spatial and temporal distributions, gear selectivity, habitat diversity and modes of invasion. To develop early detection strategies, we sampled fish in the Duluth-Superior harbor/SLR. Our extensive and comprehensive sampling (2006-2007) provides data on the distribution and abundance of known invasive, and other “rare” but native fish species, and was designed to evaluate different sampling strategies. The study used two site allocation designs (spatially balanced and targeted) and four gears (electrofishing, fyke net, bottom trawl, and beach seine). Both designs performed similarly, suggesting that targeted sampling at fewer sites may achieve the same results as spatially broader random sampling, provided the range of available habitats is included. Individual species detection probabilities and the rate of additional species acquisition varied among gears, and no single gear captured all species. Electrofishing was most efficient in terms of new species per number of fish captured, while “rare” species were captured most frequently by electrofishing or fyke netting. Total richness was generally highest in electrofishing or fyke nets, but varied spatially. Our research indicates that a complex system such as SLR could be effectively sampled with approximately 30 sampling events given appropriate allocation of effort among habitats, sites and gears.