INFLUENCES OF WATERSHED CHARACTERISTICS ON TOTAL AND METHYL MERCURY LEVELS IN LAKE MICHIGAN TRIBUTARIES

James P. Hurley^1,2,3, Kristofer R. Rolfhus³, Susan E. Cowell³, Martin M. Shafer³, and Peter E. Hughes⁴.

¹University of Wisconsin Water Resources Institute, 1975 Willow Dr., Madison, WI 53706 (hurley@wri.wisc.edu); ²Bureau of Research, Wisconsin Department of Natural Resources, 1350 Femrite Dr., Monona, WI 53176; ³Water Chemistry Program, University of Wisconsin, 660 North Park St., Madison, WI 53706; ⁴U.S. Geological Survey, 8505 Research Way, Middleton, WI 53562.

ABSTRACT

As a part of an overall project evaluating lake-wide contaminant transport, the Lake Michigan Mass Balance Project (LMMBP), we investigated total mercury (Hg_T) loadings via tributaries to the lake. Full-scale tributary sampling was conducted during two hydrologic years, from March 1994 through October 1995. During 1995, all study eleven tributaries were also sampled for methylmercury (MeHg). Mean unfiltered Hg_T concentrations in 1995 ranged from 1 ng L^-1 in the Muskegon River, to about 10 ng L^-1 in the Indiana Harbor Ship Canal and Kalamazoo River to >25 ng L^-1 in the Fox River, consistent with 1994 data. Mean unfiltered methyl mercury (MeHg) ranged from about 0.05 ng L^-1 in the Indiana Harbor Ship Canal to > 0.25 ng L^-1 in the Sheboygan and Menominee Rivers. Our initial estimates for methylmercury loading via tributaries in 1995 indicates that riverine MeHg trends are quite different from those of Hg_T. We observed significantly greater relative inputs of MeHg from watersheds with a large proportion of forest and wetland areas, an observation consistent with our “Background Trace Metals in Wisconsin Rivers” study (Hurley et al. 1995). Methyl Hg bioaccumulation in Lake Michigan, then may be enhanced in mixing zones from rivers considered “remote” rather than those with direct anthropogenic discharges.

INTRODUCTION

Transport by rivers is an important vector for mobilization of Hg in the Great Lakes Basin (Babiarz and Andren 1995; Hurley et al. 1995; 1996; 1998a,b; Mason and Sullivan 1997). In addition to receiving geologic and atmospheric inputs of Hg, tributaries of the Great Lakes typically serve as receptors for industrial and municipal discharges (Glass et al. 1990; Hurley et al. 1995; 1996; 1998a,b. In the Lake Michigan watershed, several tributaries and associated harbors are currently listed by the US Environmental Protection Agency as “Areas of Concern”. The estuarine zones of these rivers may serve as important zones of biotic production and reproduction. These zones then, represent sensitive areas and the quantity and partitioning behavior of a contaminant can be ultimately linked to biotic quality.

Here, we discuss riverine concentrations and fluxes of both Hg_T and MeHg in eleven study tributaries of the LMMBP during 1995. Since inorganic Hg has been shown to be converted to MeHg in watersheds, we assess the importance of specific tributaries in delivering the bioaccumulative form of Hg (MeHg) to nearshore regions of Lake Michigan.

METHODS

	Drainage Area (km^2)	Land Use/Land Cover				Samples (n)	Mean Conc
Tributary		Urban	Agric	For.	Wet.		HgT (ng L^-1)	MeHg (ng L^-1)	% as MeHg
Fox (Lower Fox)	16,429 (1,135)	2.8 (16.9)	52.0 (76.2)	26.4 (3.5)	13.0 (0.7)	18	21.9	0.162	0.74
Grand	14,395	5.5	75.5	13.9	3.7	21	3.82	0.111	2.91
Indiana Harbor Ship Canal	179	78.2	1.5	3.0	6.2	9	10.23	0.050	0.48
Kalamazoo	5,125	6.1	75.1	12.6	4.2	17	10.94	0.150	1.37
Manistique	3,798	0.3	5.1	50.1	39.7	12	3.88	0.119	3.07
Menominee	10,556	0.7	7.1	73.0	16.4	12	4.02	0.242	6.02
Milwaukee	2,259	21.7	66.2	7.0	3.7	21	4.51	0.150	3.33
Muskegon	7,092	2.8	33.4	47.8	11.2	12	0.98	0.086	8.78
Pere Marquette	1,775	0.7	17.2	71.4	8.4	14	2.42	0.112	4.63
Sheboygan	1,089	3.8	80.7	6.6	7.6	21	5.45	0.360	6.61
St. Joseph	12,155	5.5	80.5	9.3	2.4	16	5.32	0.106	1.99
Table 1. Watershed characteristics, HgT and MeHg in Lake Michigan tributaries during 1995. Values of watershed components for the Fox River are expressed both as an entire watershed and the Lower region from Lake Winnebago to Green Bay.

Eleven Lake Michigan tributaries chosen for this study drain watersheds with a wide range of land use/land cover types ranging from highly industrial (Indiana Harbor Ship Canal) to wetland/forest (Menominee, Manistique R.) systems (Table 1). As a result, tributaries reflected a wide range of levels of suspended particulate matter and dissolved organic carbon levels (Hurley et al. 1998b). A flow-stratified sampling strategy was adopted of which about 40% of our samples were obtained during high flow events (exceedence of historic 20% flow probability).

Ten of the eleven rivers selected for study were considered Areas of Concern by USEPA due to past contamination in the harbor or upstream reaches of the tributary. One site, the Pere Marquette River, was chosen as a “background” river, and the sampling site was located upstream rather than at the mouth of the river. Sampling teams from the U.S. Geological Survey, Wisconsin Department of Natural Resources and the Michigan Department of Natural Resources followed strict trace metal clean sampling procedures. Details of the sampling and analytical protocols are given elsewhere (Hurley et al. 1995; 1998a,b). Generally, clean sampling techniques were used during all stages of bottle preparation, sampling and analysis following accepted techniques for Hg sampling and analysis (Patterson and Settle 1976; Fitzgerald and Gill 1979; Fitzgerald and Watras 1989; Bloom and Crecelius 1983; Horvat et al. 1993; USEPA 1995).

RESULTS AND DISCUSSION

Mean concentrations of Hg_T during 1995 (Table 1) were similar to those measured during 1994 of the LMMBP (Hurley et al. 1996; 1998a,b). Highest concentrations of Hg_T were observed in the urban and industrially dominated Fox River in Wisconsin (22 ng L^-1) while the lowest in the Muskegon River (1 ng L^-1). The depositional basin of Lake Muskegon upstream of the sampling site, is assumed to scavenge particle-bound Hg.

In contrast to Hg_T, MeHg were not highest in the anthropogenically-influenced sites. In the Fox River, MeHg levels were slightly elevated (0.162 ng L^-1) relative to other sites, and the lowest mean MeHg was found at the Indiana Harbor Ship Canal. These results suggest that despite the fact that Hg_T concentrations were elevated, the form of Hg in these contaminated sites was not available for transformation to the bioaccumulative MeHg form. The highest mean MeHg were observed in the Sheboygan and Menominee Rivers. The increase in the Sheboygan River is consistent with our studies in 1992 and 1993 (Hurley et al. 1995) that showed a significant increase in MeHg below the Sheboygan Marsh in the upper reaches of the river. The Menominee River drains a watershed with a significant forest and wetland component. It is also interesting to note that the Pere Marquette River, consider “background” and uncontaminated for our study, was about mid-range for mean MeHg concentration, but due to lower Hg_T levels, MeHg accounted for almost 5% of the mercury in the river. The greatest percentage of Hg_T as MeHg was observed in the Muskegon River, downstream from Lake Muskegon. It is not known whether Hg_T is more efficiently scavenged within the lake, or if the lake itself was a source of MeHg.

Clearly these results suggest differing reactivities of Hg_T in contrasting watersheds. It has been shown that wetlands are significant sites of Hg methylation (St Louis et al. 1995; Hurley et al. 1995; Gilmour et al. 1998) however; our results also suggest that forested regions can also contribute to MeHg inputs. The Pere Marquette River watershed is about 71% forest and only 8% wetland and the mean MeHg concentration is similar to the Manistique River where wetlands constitute 40% of the surface area. In addition to using simple watershed coverages to explain Hg dynamics, it is most likely important to consider hydrologic flow paths and connectivity of specific land cover types to the main river channel.

Using instantaneous flow data from USGS gages or acoustic velocity meters, it is possible to calculate both total and methyl Hg inputs to the lake via the study tributaries (Figure 1). These results reveal that on a loading basis, the Fox River is the most important tributary for input of total and methyl mercury. However, the Menominee River, a river with an order of magnitude lower Hg_T loading than the Fox River, delivers a comparable amount of MeHg to Lake Michigan. Both of these tributaries directly discharge into Green Bay rather than the open reaches of the lake. It is also interesting to note that despite elevated Hg_T levels in the Indiana Harbor Ship Canal, where particles contain over 1 mg g^-1 Hg on a dry weight basis (Hurley et al. 1998a), relatively little MeHg is produced. Taken together, the above results suggest substantially different reactivities of inorganic Hg among tributaries. Therefore, in mixing regions of forest and wetland dominated tributaries one may observe enhanced bioaccumulation of Hg in Lake Michigan.

ACKNOWLEDGMENTS

This work was supported by the U.S.E.P.A. Region V under Grant No. X995469010 to the Wisconsin Department of Natural Resources and the University of Wisconsin. We thank the field sampling teams from the U.S.G.S., the W.D.N.R. and the M.D.N.R. for their care and patience during sampling. S. Byler, R. Baldino, T. Heelan, S. Claas and J. Knepel provided laboratory assistance. D. Robertson (USGS-Middleton, WI) provided GIS data of watershed coverages.

REFERENCES

Babiarz, C.L., Andren A.W. (1995) Water Air Soil Pollut. 83:173-183.

Bloom, N.S., Crecelius E.A. (1983) Mar. Chem.14:49-59.

Fitzgerald, W.F., Gill, G.A. (1979) Anal. Chem.51: 1714-1720.

Fitzgerald, W.F., Watras C.J. (1989) Sci. Tot. Environ.28:223-232.

Glass, G.E., Sorensen, J.A., Schmidt, K.W., Rapp G.R (1990) Environ. Sci. Technol. 24:1059-1069.

Gilmour, C.C., Ederington, M.C., Bell, J.T., Gill G.A., Stordal M.C. (1998) Biogeochem. 40:327-345.

Horvat, M., Bloom, N.S., Liang, L.(1993) Anal. Chim. Acta 282:153-168.

Hurley, J.P., Benoit J.M., Babiarz C.L, Shafer M.M., Andren A.W., Sullivan J. R., Hammond R., Webb D.A. (1995) Environ. Sci. Technol. 29:1867-1875.

Hurley, J.P., Shafer M.M., Cowell S.E., Armstrong D.E., Overdier J.T., Hughes P.E. (1996) Environ. Sci. Technol. 30:2093-2098.

Hurley, J.P., Cowell S.E., Shafer M.M., Hughes P.E. (1998a) Environ. Sci. Technol. 32:1424-1432.

Hurley, J.P., Cowell S.E., Shafer M.M., Hughes P.E. (1998b). Sci. Tot. Environ. 213: 129-137.

Mason, R.P., Sullivan K. (1997) Environ. Sci. Technol., 31 (3), 942 –947.

Patterson, C.C. and D.M Settle (1976) In, Accuracy in Trace Analysis: Sampling, Sample Handling, and Analysis. P.D. LaFleur, Ed., U.S. National Bureau of Standards Special Publication 422, pp.321-351.

St. Loius, V.L., Rudd J.W.M., Kelly C.A., Beatty K.G., Flett R.J., Roulet N.T. (1996) Environ. Sci. Technol. 30:2719-2729.

U.S. Environmental Protection Agency. 1993. Lake Michigan Tributary Monitoring Project Plan. USEPA Region V, Chicago, IL.

U.S. Environmental Protection Agency. 1995. Method 1631: Mercury in water by oxidation, purge and trap and cold vapor atomic fluorescence spectrometry. Office of Water, EPA 821-R-95-027.