Historical Trend of Mercury Deposition in Seneca Lake, NY

Mercury (Hg) contamination is pervasive in aquatic ecosystems and its bioaccumulation may lead to severe health concerns for both wildlife and humans. In this study we quantify historical Hg fluxes to the profundal sediment of Seneca Lake, NY. Previous studies show that the highest surface sediment Hg concentrations occur here, near a major tributary, the Keuka Lake Outlet. This tributary hosted much of the early industry in the area. Analysis of a 137Cs and 210Pb-dated sediment box core indicates total Hg (HgT) fluxes were low (197 μgm-2y-1) in 1770 and peaked between 1890 and 1910 (583 μgm-2y-1) and gradually returned to regional background levels (127 μgm-2y-1) by 1977. This peak in HgT flux predates those observed in other local and regional lakes, suggesting that a point source rather than widespread atmospheric deposition is the reason for increased HgT flux to the sediment. Early industry, including tanneries, hatteries, paper mills, and a flourishing nursery market, as well as a growing population during the late 19th century are possible sources of the high Hg concentrations found in Seneca Lake sediment. Introduction. Mercury (Hg) is a toxic trace metal pervasive in aquatic ecosystems across the Northern Hemisphere (Fitzgerald and Clarkson, 1991). Its bioaccumulation may lead to severe health concerns for wildlife and humans (US EPA, 1997). Mercury, in its elemental form, has a relatively long atmospheric residence time of up to three years (Perry et al., 2004). Highly volatile, Hg can spread from the atmosphere to the hydrosphere and biosphere quickly, and is known to cause neurological problems in vertebrates. Due to its bioavailability, mercury bioaccumulates throughout the food chain, transferring methylated forms from contaminated waters into fish and then to higher mammals and humans. Since 2001, sixty-three New York State (NYS) lakes, reservoirs, and ponds have been added to the Department of Health’s fish consumption advisory list due to elevated levels of mercury (Callinan, 2001). Natural (e.g., forest fires, volcanic eruptions) and anthropogenic (e.g., fossil fuel combustion, medical waste incineration, municipal waste combustion, metal smelting, coal combustion) processes may supply Hg to the environment (US EPA, 1997). At regional to global scales, the primary mechanism of Hg contamination is through atmospheric transport and subsequent deposition via dust, rain, snow (US EPA, 1997). At the local scale, Hg may enter aquatic ecosystems through direct point sources (Bookman et al., 2008). Previous study of sediment box cores from four lakes in central NYS found that the source of Hg differed spatially and temporally (Bookman et al., 2008). Three of the four lakes investigated showed peaks in Hg deposition (0.9 ppm) during the 1970s as a result of a significant regional atmospheric source (Bookman et al., 2008). The fourth lake showed a peak in Hg deposition (0.16 ppm) during the 1990s, which may be associated with a new, local emission source. In order to assess the sources of Hg to the New York Finger Lakes, Abbott and Halfman (2009) analyzed the Hg concentration of surface dredge samples from Seneca Lake, the largest Finger Lake. The average concentration (0.127 ppm) in Seneca Lake is high compared to the other four NYS, suggesting both regional and local sources both contribute mercury. Based on a survey of dredge samples from Seneca Lake in 1975, Blackburn, Fogg and Cornwall (1979) identified two potential local point sources: the local coal-fired power plant in Dresden, NY and a defunct industry in the Keuka Lake Outlet watershed. To test this hypothesis and evaluate the relative importance of local versus regional sources of Hg to the Finger Lakes region over time, we analyzed the Hg content of a sediment box core from Seneca Fig. 1. Box core location in Seneca Lake NY. Lake, at a high resolution (1 cm; 6-7 years). A better understanding of the long-term history of atmospheric Hg cycling and quantification of the importance of local and regional sources of Hg is critical for assessing the impact of anthropogenic Hg emissions and to develop effective emission control strategies. Study Area. The 11 Finger Lakes of central New York occupy deep, narrow, north-south oriented glacially scoured basins (Schaffner and Oglesby, 1978). The largest by volume (15.54 km3), Seneca Lake has a surface area of 175 km2, and is 57 km long, 5.2 km wide at its maximum and up to 188 m deep (Schaffner and Oglesby, 1978). The U-shaped lake has steep bedrock walls composed of Paleozoic sedimentary rocks (Schaffner and Oglesby, 1978). Soils of the region are Spodosols and Gleysols that developed in Quaternary glacial and glaciolacustrine deposits (USDA, 1958). Of Seneca Lake’s of 1181 km2 watershed, 41% of the watershed is forested, 54% is devoted to agriculture, and the remainder, residential and urban (Schaffner and Oglesby, 1978). The climate of Geneva, NY is typical of a midlatitude, temperate continental site, with cold, snowy winters and relatively mild, humid summers (Schaffner and Oglesby, 1978). Average temperatures range from 3.6°C in January to 26.5°C in July and the mean annual precipitation is 80 cm (Schaffner and Oglesby, 1978). Approximately 75% of the water reaching Seneca Lake comes from spring snowmelt (Michel and Kraemer, 1995). There are 16 major streams that enter the lake but only one outlet, located at the northeast corner of the lake (Fig. 1). Another Finger Lake, Keuka, drains into Seneca Lake via the Keuka Lake Outlet, and accounts for ~30% of Seneca Lake’s watershed. Seneca Lake is hydrologically open and the residence time is between 12-18 years (Michel and Kraemer, 1995). Seneca is an oligo-mesotrophic, warm monomictic lake. Lake waters are well-mixed annually, alkaline (pH = 8.1), and characterized by high concentrations of Cl(139 mg/L), SO4 (38 mg/L), alkalinity (106 mg/L, as CaCO3), Na+ (79 mg/L), Ca2+ (42 mg/L), Mg2+ (11 mg/L), and K+ (2.7 mg/L) (Halfman et al., 2006). Waters remain well-oxygenated year round. Sampling. Box Core Collection and Sampling. A 36-cm long sediment core were collected from the profundal zone of Seneca Lake at 113 m water depth, north of the present day town of Dresden (42°46.296’N, 76°56.872’W) and between two streams, Kashong Creek and Keuka Lake Outlet using a modified Wildco box corer (Fig. 1). This site was selected because it had one of the highest concentrations of HgT at the sediment-interface in 1975 (Blackburn et al., 1979) and in 2008 (Abbott and Halfman, 2009). The core remained upright to preserve the sediment-water interface until sectioned. The core was extruded and sampled at a 1-cm interval at Hobart & William Smith Colleges (HWS). The outer smear was removed from each sediment increment with a stainless steel spatula, and samples were transferred into plastic containers. Samples were stored at 4°C until processing. Subsamples were sent to Rensselaer Polytechnic Institute for 210Pb and 137Cs analyses. The remainder of the sediment was frozen, freeze-dried and mechanically homogenized in plastic bags. Freeze-dried sediments were stored at room temperature. Analytical Methods. Sediment Dating and Chronology. Radiochemistry analyses of 137Cs and 210Pb at the Rensselaer Polytechnic Institute were conducted to develop an age chronology. 137Cs is associated with global fallout derived from atmospheric testing of nuclear weapons beginning in the early 1950s, with a fallFig. 2. 137Cs peak versus depth in box core S3. A worldwide peak in 137Cs occurs at 1963-64. out maximum in 1963-1964. The activity of 210Pb, a naturally occurring radionuclide derived from the decay of atmospheric radon, decreases exponentially from the surface of the sediment with a half-life of 22.3 years. An age-depth model was constructed using a combination of 137Cs and 210Pb data. After identifying the Cs peaks, a regression line was fit mathematically to the 210Pb data. The slope of this line was used to calculate sediment accumulation rates and extrapolate ages for the core. Loss-on-Ignition. Water, organic matter and carbonate content of dry sediment was determined on ~1 g using standard loss-on-ignition techniques (LOI) (Dean, 1974). Subsamples were first freeze-dried to determine the weight percent water content. The same subsamples were analyzed for weight percent total organic matter and carbonate content by LOI at 550°C and 1000°C, respectively. Mass measurements were made before and after freeze-drying and each heating using an electronic analytical balance. Dry density was calculated from water content, and fixed densities for the organic, carbonate and inorganic fractions. Grain Size. Subsamples were treated with hydrogen peroxide and glacial acetic acid to remove the organic matter and carbonate fractions, respectively, and isolate the terrigenous component of the sediment using the methods of (Jackson, 1969). Samples were analyzed using a Coulter LS 230 Multivariable Laser Diffraction Particle Size Analyzer in duplicate. Grain size statistics were calculated using Folk and Ward’s (1975) equations. Mercury. Total Hg concentration (HgT) of the box core was determined at the USGS Woods Hole Field Station using a direct mercury analyzer (DMA-80, Milestone Inc. Monroe, CT USA), which uses thermal decomposition, gold amalgamation and atomic absorption spectrometry and has a detection limit of 0.009 ng. Between 0.04 and 0.07 mg of each sample was analyzed in duplicate with one in six samples run in triplicate to calculate standard deviation. A calibration curve was generated using two reference materials: National Research Council of Canada Institute for National Measurement Standards MESS-3 (marine sediment, certified value = 91 ± 9 ng/g HgT [dry weight]) and US Commerce Department National Institute of Standards and Technology SRM 1515 Apple Leaves (certified value = 44 ± 4 ng/g HgT [dry weight]). Machine blanks were analyzed every 5-10 samples and reference samples were analyzed every 5-20 samples. Fluxes of HgT to the lake were calculated as the product of the 210Pb-derived sedimentation rates, sediment density, and dry weight HgT concentrations for each core interval and reported as mg m-2 y-1. Results. Dating. The 137Cs is associated with global fallout derived from atmospheric testing of nuclear weapons beginning in the early 1950s, with a fall-out maximum in 1963-1964 (Fig. 2). Based on the 137Cs peaks, the net accumulation rate is approximately 0.2 cm y-1 between 0 and 12 cm depth in the core. The activity of 210Pb, a naturally occurring radionuclide derived from the decay of atmospheric radon, decreases exponentially from the surface of the sediment with a halflife of 22.3 years. The core exhibits an exponential decline in 210Pb with 1953 1963-64