Optimizing Stream Water Mercury Sampling for Calculation of Fish Bioaccumulation Factors

Abstract

Mercury (Hg) bioaccumulation factors (BAFs) for game fishes are widely employed for monitoring, assessment, and regulatory purposes. Mercury BAFs are calculated as the fish Hg concentration (Hg(fish)) divided by the water Hg concentration (Hg(water)) and, consequently, are sensitive to sampling and analysis artifacts for fish and water. We evaluated the influence of water sample timing, filtration, and mercury species on the modeled relation between game fish and water mercury concentrations across 11 streams and rivers in five states in order to identify optimum Hg(water) sampling approaches. Each model included fish trophic position, to account for a wide range of species collected among sites, and flow-weighted Hg(water) estimates. Models were evaluated for parsimony, using Akaike's Information Criterion. Better models included filtered water methylmercury (FMeHg) or unfiltered water methylmercury (UMeHg), whereas filtered total mercury did not meet parsimony requirements. Models including mean annual FMeHg were superior to those with mean FMeHg calculated over shorter time periods throughout the year. FMeHg models including metrics of high concentrations (80th percentile and above) observed during the year performed better, in general. These higher concentrations occurred most often during the growing season at all sites. Streamflow was significantly related to the probability of achieving higher concentrations during the growing season at six sites, but the direction of influence varied among sites. These findings indicate that streamwater Hg collection can be optimized by evaluating site-specific FMeHg-UMeHg relations, intra-annual temporal variation in their concentrations, and streamflow-Hg dynamics.

Optimizing fish sampling for fish–mercury bioaccumulation factors

“…The USEPA (U.S. Environmental Protection Agency, 2010) allows for calculation of BAF based on MeHg or total Hg in water (empirical comparison provided in Riva- Murray et al, 2013a). Although water is the denominator in the BAF equation, it is understood that diet is the dominant mechanism of fish MeHg uptake (Hall et al, 1997).…”

“…Herein, stream-water MeHg concentration and top-predator-fish Hg concentration, age, size, and base-adjusted trophic position data collected using a standardized protocol from 11 streams spanning large gradients in environmental setting and atmospheric mercury deposition across the United States during 2003-2009 are employed to address this information gap. Optimization of BAF Hg water sampling methodologies (such as collection, timing, filtration, and analysis of MeHg versus THg) is addressed elsewhere (Riva-Murray et al, 2013b).…”

“…However, a wider range of sizes was collected from most sites to obtain sufficient numbers of specimens. Fish trophic positions were estimated from d 15 N that was adjusted (Kidd et al, 1995;Anderson and Cabana, 2007) for siteto-site variation in basal d 15 N by subtracting mean d 15 N from periphyton samples collected from selected habitats at each site, as described elsewhere Scudder, 2004, 2007;Beaulieu et al, 2012;Riva-Murray et al, 2013b, 2013c. Periphyton was used instead of a common lowest trophic-level consumer , because no single primary consumer species/feeding guild was available across all sites (Beaulieu et al, 2012;RivaMurray et al, 2013b).…”

“…Samples for NY and SC were analyzed for d 15 N at the Stable Isotope Geochemistry Laboratory (Florida State University, Tallahassee, FL) with a ThermoQuest NC2500 Elemental Analyzer interfaced with a Finnigan MAT Delta Plus XP isotope ratio mass spectrometer. Precision and accuracy for isotopic ratios were less than 0.2 per mil for d 15 N. Results, collection and analytical method details, and quality control data for Hg water are provided elsewhere Journey et al, 2012;Riva-Murray et al, 2013b).…”

“…At each site, Hg fish concentrations were compared to mean and 90th percentile concentrations of MeHg in filtered water (FMeHg) calculated over a period of 365 d immediately preceding fish sampling. FMeHg statistics were calculated using daily FMeHg concentrations derived from rating-curve analyses that account for variation in FMeHg as a function of streamflow and season Bradley et al, 2013;Riva-Murray et al, 2013b). For Lookout Creek in OR (LKT OR ), the 90th percentiles for FMeHg were set at the MDL of 0.04 ng L À1 because 40 of 43 measurements were less than the MDL and two concentrations were at the MDL.…”