USGS Home
Contact USGS
Search USGS

 Go to:      NAWQA Home     National-Synthesis Assessments

Trace Elements National Synthesis Project

2000 conference paper--
A National Pilot Study of Mercury Contamination of Aquatic Ecosystems along Multiple Gradients: Bioaccumulation in Fishes

By William G. Brumbaugh, David P. Krabbenhoft, Dennis R. Helsel, and James G. Wiener

Presented at the 21st annual meeting of the Society of Environmental Toxicology and Chemistry (SETAC), Nashville, TN, November 12-16, 2000

Abstract

A national pilot study to examine relations of mercury (Hg) and methylmercury (MeHg) in aquatic ecosystems was conducted by sampling water, sediment, and fish in the summer and fall of 1998 at 106 stations from 20 U.S. watershed basins (Fig. 1). Mercury bioaccumulation in fishes was strongly (positively) correlated with the MeHg concentration in water but only moderately with the MeHg in sediment or the total Hg in water. Of the other measured parameters, pH, dissolved organic carbon (DOC), sulfate, sediment loss on ignition (LOI), and the percent wetlands of each basin were also significantly correlated with Hg bioaccumulation in fishes. The best model for predicting Hg bioaccumulation included MeHg in water, pH of the water, % wetlands in the basin, and the acid-volatile sulfide (AVS) content of the sediment. Based on rankings by various Hg criteria, sampling sites from the following five study units had the greatest Hg contamination: Nevada Basin and Range, South Florida Basin, Sacramento River Basin (California), Santee River Basin and Coastal Drainages (South Carolina), and the Long Island and New Jersey Coastal Drainages.

Figure 1. The 20 USGS NAWQA basins sampled in this pilot study, categorized by geometric mean of Hg in fish muscle (mg/kg wet wt.)

Introduction

Methylmercury (MeHg) is a potent neurotoxin that is perhaps the most widespread contaminant affecting our Nation's aquatic ecosystems. Human fish-consumption advisories for Hg in fish now include some 40 states and account for more than eighty percent of all such advisories in the Nation (USEPA, 1996, 1998). In this pilot study, bioaccumulation rates of mercury in fish axial muscle (the dominant repository for methylmercury in fish) for 20 nationwide basins were evaluated in conjunction with measures of total and methylmercury in sediment and water, and selected watershed chemical and physical characteristics. The study was conducted through the collaborative efforts of the U.S. Geological Survey's National Water Quality Assessment (NAWQA) program, its Toxic Substances Hydrology program's mercury work, its Wisconsin District Mercury Laboratory (WDML), and its Columbia Environmental Research Center (CERC). The overall objective of this study was to identify ecosystem characteristics that favor the production and bioaccumulation of MeHg on both a regional and national basis.

Study design

• Sampling of fish, water, and sediment from June to October, 1998 from 20 NAWQA basins (Fig. 1)
• 3-5 sites per basin (mostly streams) representing gradients of mercury load and watershed characteristics
• Age-3 largemouth or other bass species targeted for Hg analysis (axial muscle, composite of 3-8 individuals of similar size if available); 65% of samples collected were bass species
• Sampling and analysis of MeHg and total Hg in water and sediment by ultra-trace protocols
• Ancillary measures--
    water: pH, DOC, sulfate
    sediment: LOI, AVS
    site: percent wetlands, land use

Results and discussion

Mercury concentrations for each basin are summarized graphically in Figures 1 and 2. Summary statistics of concentrations (ug/g wet wt.) for all fish samples (n=159) were as follows: arithmetic mean, 0.478; geometric mean, 0.218; median, 0.206; minimum, 0.018; maximum, 5.84. For largemouth bass (n=50) these same statistics were 0.510, 0.329, 0.292, 0.045 and 4.22; for smallmouth bass (n=37) the values were 0.244, 0.195, 0.205, 0.042 and 1.05, respectively. Individual sites where one or more fish samples had Hg concentrations above the 0.5 ug/g wet advisory level are listed in Table 1.

Figure 2. Geometric mean of Hg concentration (ug/g wet wt.) in fish muscle samples collected from each of the 20 basins.
  ^*NVBR samples collected from only one site (Lahontan Reservoir).

Table 1. Fish samples from this study with Hg conc. > 0.5 ug / g wet wt. ( -- = no data).

A summary of the correlation of each measured variable (transformed as necessary to meet linearity requirements) with length-normalized Hg concentrations in fish is given in Table 2. The relative strength for correlation with measures of Hg in water and sediment was:
    MeHg _water > MeHg _sediment > Hg-Total _water >> Hg-Total _sediment (not significant).
With the exception of DOC, all variables that were correlated with bioaccumulation exhibited stronger relationships for largemouth bass than for all species combined. The following 4-variable model was deemed most useful (accounting for 45% of the variability) for predicting bioaccumulation rates of Hg in fish for the 20 basins in this nationwide pilot study:
    ln(Hg/length) = - 3.6 + 0.41 ln(MeHg _water) + 0.02 (%wetland) - 0.27 pH - 0.12 (AVS^1/3)

Table 2. Correlation of measured parameters with Hg bioaccumulation rate.

Geometric means of the five Hg measures when categorized according to the predominant land use for each site are depicted in Fig. 3. All land-use categories were significantly different with respect to Hg _fish. The ranking of land use with respect to Hg _fish was:
    A/F >> Mine > Ag > Urb.~ Bkd.
The greatest discrepancy between Hg in sediments and in fish was for urban watersheds, which ranked comparatively low for Hg _fish but high for Hg-Total and MeHg in sediment. Bioaccumulation factors (BAFs) for the transfer of MeHg from water to fish for each land-use category (and the South Florida basin) are plotted in Fig. 4. For largemouth bass viewed separately, BAFs appeared to be inversely related with the concentration of MeHg in water, suggesting that low concentrations of MeHg in water are more efficiently biotransferred than are high concentrations.

Figure 3. Geometric mean of total or methyl mercury in fish (ug/g wet), water (ng/L), and sediment (ng/g dry) for sites categorized according to predominant land use. Number of observations = 13, 42, 23, 15, and 34, for mixed agriculture/forest (A/F), Mine-impacted (Mine), Agriculture (Ag.), reference/background (Bkg.), and Urban (Urb.), respectively. Excludes South Florida basin.

Figure 4. Geometric mean of MeHg bioconcentration factor (Hg _fish / MeHg _water) for each land-use category: all species or largemouth bass. Units are L/kg = ng/kg(fish) divided by ng/L(water).

Summary

Widespread mercury contamination of our Nation's waterways remains a serious problem. Sources and accumulation rates in fishes vary widely among regions. One or more fish samples from 9 of the 20 basins examined in this pilot study had Hg concentrations above the advisory level of 0.5 ug Hg/g wet wt. Based on rankings of selected water, sediment, and fish criteria, sampling sites from the following five basins exhibited the greatest Hg contamination: Nevada Basin and Range, South Florida, Sacramento Basin (CA), Santee Basin and Coastal Drainages (SC), and the Long Island and New Jersey Coastal Drainages. The concentrations of Hg in fish were strongly correlated with MeHg in water, but only moderately with MeHg in sediment or Hg-Total in water. There was no correlation of Hg in fish with total Hg in sediment.

From a national perspective of flowing waters, the concentration of MeHg in water was by far the most useful variable for predicting the Hg bioaccumulation rate in fish (concentration of Hg in muscle divided by the fish length). However, the percent wetlands (+) , the pH of the water (-) , and the sediment AVS (-) also contributed significantly to a predictive model. Based on ranking criteria, sub-basins categorized as mixed agriculture/forest and mining-impacted exhibited the most consistent contamination of mercury in all three sample matrices (water, sediment, and fish). The greatest discrepancy in rankings of Hg in fish and in sediments was for urban watersheds, where sediments often ranked high but fish usually ranked low. For largemouth bass the bioaccumulation factor or BAF (Hg _fish / MeHg _water) was lowest for land-use categories having the highest concentrations of MeHg in water, suggesting that low concentrations of MeHg in water are more efficiently biotransferred than high concentrations. Nevertheless, high concentrations of MeHg in water will generally yield high concentrations in fish.

Acknowledgments

The authors are indebted to all NAWQA study-basin teams who participated and demonstrated great diligence, commitment, and enthusiasm in the collection of the samples. We thank Jude Smith and coworkers with the U.S. Fish and Wildlife Service in Albuquerque, NM, for conducting age determinations and Dr. Ted Lange (Florida Fish and Game) for providing Hg measures in fish from South Florida. We also acknowledge the technical assistance of our colleagues Jesse Arms, Mike Walther, and Mark Olson.

References

USEPA, 1996, National Listing of Fish and Consumption Advisories Database: United States Environmental Protection Agency, EPA-823-C-96-011, Office of Water, Washington , DC, 7 diskette set.

USEPA, 1998, Update: Listing of Fish and Wildlife Advisories: United States Environmental Protection Agency, LFWA Fact Sheet EPA-823-98-009, Office of Water, Washington , DC, 4p.