Spatial variations in fish-tissue mercury concentrations in the St. Croix River basin, Minnesota and Wisconsin, 2004

Parts of the St. Croix River in Minnesota and Wisconsin are under fish-consumption advisories because of elevated mercury concentrations that have been measured in fish from this river. The U.S. Geological Survey, National Park Service, and the University of Wisconsin, LaCrosse, cooperated in a study to determine the spatial variation of mercury in fish in the St. Croix River and selected tributaries.

Game and nongame fish were collected at 14 sites during summer 2004 and identified to species. One hundred ninety-three (193) composite tissue samples were analyzed for total mercury as whole fish, skin-on fillet, or skin-off fillet. A model of mercury in fish was used to standardize fish-tissue mercury concentrations to a common species, tissues sampled, and length of fish allowing for more consistent comparisons among sites.

Rush Creek near Rush City, Minnesota, was identified as having high median standardized fish-tissue mercury concentrations compared to other tributaries sampled. Previous studies identified Rush Creek as having high concentrations of methylmercury in water and high concentrations of total mercury in sediment when compared to other sites in the St. Croix River Basin.

Sites in the St. Croix River Basin that drained forest/wetland watersheds had significantly higher median fish-tissue mercury concentrations than sites draining agricultural/forested watersheds (p=0.0003). There also was a significant relation between fish-tissue mercury concentration and methylmercury concentration in water (rho=0.580, p=0.02) and between fish-tissue mercury and total mercury in sediment (rho=0.569, p=0.03). Observed fish-tissue mercury concentrations exceeding the U.S. Environmental Protection Agency’s (USEPA) human-health criterion of 300 micrograms per kilogram occurred at 7 of the 14 sampling sites. The model predicted concentrations exceeding USEPA’s criterion at all of the seven sites where exceedances were observed and four of the seven sites where exceedances were not observed. The implication is that fish-consumption advisories that are based on observed concentrations (of a subset of the species that occur at the site or smaller fish) could underestimate the threat to human health.

Using the model to predict fish-tissue mercury concentrations allows site-specific fish-consumption advisories to be developed for multiple species and different lengths of fish. Potential mercury exposure to fish consumers may be reduced because an individual can choose to consume sizes and species of fish that are expected to have lower fish-tissue mercury concentrations. The National Park Service can use these results to more reliably monitor fish-tissue mercury concentrations in the St. Croix River Basin and better assess potential health effects of fish consumption to humans and wildlife.