Hg and other TRACE
metals in water, suspended matter, and biota in the Chattahoochee River
Klaus Neumann*, W. Berry Lyons (Department of Geology, University of Alabama, Tuscaloosa, AL 35487-0338. Current address: Byrd Polar Research Center, Ohio State University, 1090 Carmack Road, Columbus, OH 43210), Emma J. Rosi, Judy L. Meyer (Institute of Ecology, University of Georgia, Athens, GA 30602)
Urban areas greatly impact surface water quality as runoff from impervious surfaces, population density and fossil fuel usage increase. Yet, little detailed quantitative data exist in order to evaluate the role of urbanization on river systems. We have collected a series of water samples from the Chattahoochee River as it flows through the Atlanta metropolitan area, and analyzed the samples for mercury, major and trace metals (e.g., V, Cu, Mo). In addition, we present data from the farther downstream, less urbanized portion of the river. The dissolved concentration of Na increases three-fold as the river proceeds through Atlanta, while the dissolved Hg concentrations increase two to four fold (from 0.15 ng/L to 0.7 ng/L). Particulate Hg varies seasonally, from less than 30% (68 to 89 ng/g) to more then ten-fold increase (2 to 200 ng/g). The increase in Hg is reflected in elevated concentrations of mercury and other metals in aquatic insects from the river. Dilution and/or removal processes decrease the maximum values observed as the river proceeds through West Point Lake, southwest of Atlanta. The samples from the lower Chattahoochee (down to the Apalachicola River in Florida) show an increase in dissolved Hg (0.12 to 3.174 ng/L), but a decrease in particulate Hg (0.18 to 0.018 µg/g). The dissolved Hg will be compared to the behavior of other trace metals in the river, which decrease in concentration downstream from Atlanta. With the possible exception of Hg, the trace metal input from metro-Atlanta is clearly seen more than 300 km downstream.
Trace metals, even at very low background concentrations, can be strongly accumulated through the foodweb in aquatic systems (i.e., the concentration increases from water to insects to fish) and, potentially, be harmful to humans through fish consumption. Therefore, the study of the biogeochemistry and bioaccumulation of trace metals, such as Hg, should be an important aspect of environmental monitoring programs. This is particularly true in urban areas where the anthropogenic emissions are greatest per unit area.
This work was undertaken to
establish the role of the Atlanta metro region on the trace metal
biogeochemistry of the Chattahoochee River. To achieve this, water samples were
taken seasonally during one year along a transect of ~100km through the center
of Atlanta. In addition, another set of samples was taken along the lower
stretch of the Chattahoochee River, in order to document the downstream impact
or “ecological footprint” of the Atlanta metropolitan area.
Methods
Ultraclean techniques and procedures were utilized to collect the river samples. Samples were collected and stored in precleaned teflon bottles. Samples were also collected for major cations and anions, dissolved organic carbon (DOC) and pH at each sampling location. Aliquots of the samples for Hg analysis were filtered through precleaned 0.47µm nylon filters for dissolved Hg (Hgdiss), while another unfiltered aliquot was used for the determination of total Hg (HgT). Both the filtered an unfiltered sample were acidified with Optima® HCl to about 1% concentration. The mercury was analyzed using a Brooks-Rand cold-vapor atomic fluorescence spectrometer, using the method of Bloom and Crecelius (1983). Particulate Hg (Hgpart) concentrations were obtained by subtracting Hgdiss from HgT, and accounting for the total suspended sediment (TSS) mass in each sample. Field blanks were routinely subtracted from measured values. The detection limit (3 times SD) was 0.06ng/L, and the coefficient of variation 5%. Aliquots of the trace metal samples were also filtered, and all samples were acidified with Optima® HNO3 to 2%. The samples were run on a Perkin-Elmer Elan 6000 inductively coupled plasma mass spectrometer. SLR3, a NRCC riverine reference material was also analyzed as an analytical check on accuracy. Major cations were determined with ICP. DOC samples were collected in acid-cleaned amber bottles, filtered, and measured with a total organic carbon analyzer.
Aquatic insects were collected in the field from snags. Teflon forceps were used and samples were placed directly in acid washed HDPE sample bottles. Insects were allowed to depurate for 24 hours and then rinsed with deionized water and frozen for storage. After freeze-drying, the insects were dissolved in Optima® HNO3 at 90şC for 1 day, and then analyzed like the water samples via ICP-MS and CV-AFS.
The Chattahoochee River upstream from Atlanta has very low Hg concentrations, which fall well within the range of what had previously been termed background levels in this area (Mastrine et al., 1999). The Hgdiss concentrations were between 0.15 and 0.49ng/L, and Hgpart concentrations were between 2 and 70ng/g. These Hgpart data are similar to the NAWQA sediment data of Couch (1997, and USGS/NAWQA websites quoted therein) for the river in this region. As the river flows through Atlanta, Hg concentrations in the water increase. At Franklin, which is located downstream from the last Atlanta sewage treatment plant, dissolved Hgdiss values range from 0.44 to 0.70 ng/L, and Hgpart from 90 to 200ng/g. HgT values vary from an average of 0.4ng/L above Atlanta to an average of1.6ng/L at Franklin, a four-fold increase as the river flows through Atlanta.
When the river leaves West Point Lake, the first reservoir downstream from Atlanta, particulate and dissolved Hg values decrease with Hgdiss (0.12ng/L) being lower than in the water feeding into the reservoir. However, downstream from West Point Lake the Hgdiss increases again. At Lake Seminole, Hgdiss concentrations reached 1.36ng/L, and at Bristol (the river is here called Apalachicola River), Hgdiss was 3.48ng/L. At the same time, the concentration of Hgpart decreases from West Point Lake (180ng/g) to Lake Seminole (20ng/g). The Hgpart data are essentially identical to NAWQA sediment data along the entire length of the river. While dissolved Hg and particulate Hg vary, total Hg concentrations remain nearly constant at 1.2-1.4ng/L all the way to Lake Seminole. This change in physico-chemical form between particulate and dissolved is accompanied by a small increase in dissolved organic carbon (DOC, 1.9-2.4mg/L). The additional DOC probably origins in wetlands and riparian areas of the coastal plain physiological province, which the Chattahoochee enters at Columbus, Ga. Since the DOC increases only very little, and the major ion composition of the water virtually remains the same, we speculate that a possible change in the DOC composition caused these variations in Hg speciation. In Lake Seminole, the Flint River, a coastal plain river, joins the Chattahoochee. NAWQA data show that this river has higher Hg concentrations in its sediments, while its DOC is comparable with the Chattahoochee. The Flint River may contribute a significant amount of Hgdiss to the system, which may explain the elevated Hgdiss found in the Apalachicola River further south.
NAWQA data show an increase in Hg in clams between Columbus, Ga., and Lake Seminole, which they could not explain solely by the sediment Hg data (which, like our suspended material data, show a decrease downstream). A change of the format of Hgdiss as it flows through this portion of the river may affect Hg bioavailability.
Hg behaves differently in the Chattahoochee than the other trace metals measured by us. Like Hg, the other dissolved metals increase, as the river flows through Atlanta. For example, Mo increases from 0.17 to 5.0µg/L, and V from 0.1 to 0.7µg/L. These seven to thirty-fold increases are much larger than the four-fold increase of Hg, but these metals decrease significantly as the Chattahoochee passes through the first reservoir, West Point Lake. Once past that reservoir, oxyanions such as V, As and Mo concentrations remain elevated at about the three-fold concentrations compared to values above Atlanta. Unlike HgT, metals such as Cu decrease throughout the entire stream length south of Atlanta, with nearly all the metal added in the Atlanta area later removed. The particulate metal concentrations also decrease as the Chattahoochee flows towards Lake Seminole.
At the time of writing this extended abstract, we only have a small amount of data on metal concentration in aquatic insects. Preliminary insect data suggest that insects in the part of the river just below Atlanta have higher metal concentrations than the insects living in upstream areas. The highest Hg concentration found was >800ng/g, which is the same magnitude as observed in zooplankton in recently flooded reservoirs in Canada (Plourde et al., 1997). It is our goal to obtain reliable Hg and metal data for the insect and fish samples within the Atlanta metropolitan area within the next few months.
In conclusion, Hg concentrations in the Chattahoochee River in the Atlanta metropolitan area increase by a factor of four, but still remain in the range of naturally occurring concentrations. When the Chattahoochee River enters the coastal plain, the physicochemical form of mercury changes from Hgpart to Hgdiss, perhaps caused by changes in the DOC contents of the river. Interestingly, the HgT concentrations change little until the confluence of the Flint River, ~300km south of Atlanta. Unlike mercury, other metals show varying degrees of removal from the water downstream of Atlanta.
We acknowledge the support of the Turner Foundation, and we greatly appreciate the energy and skill of Ms. EY Graham in the collection and analysis of these data.
Bloom NS, Crecelius EA (1983), Mar. Chem. 14: 49-59
Couch CA (1997), Proc. 1997 Georgia Water Res. Conf., U. of Georgia, pp. 45-52.
Mastrine JA, Bonzongo J-CJ, Lyons WB (1999), Appl. Geochem. 14: 147-158
Plourde Y, Lucotte M, Pichet P (1997), Can. J. Fish. Aquat. Sci. 54: 821-831