James A. Barres

Trace Element Concentrations in Near-Surface Waters of the Great Lakes and Methods of Collection, Storage, and Analysis

Abstract From 1980 through 1985, waters of the Great Lakes were sequentially sampled for dissolved, paniculate, and total trace elements. Major sampling occurred in 1980 for Lake Huron, in 1981 for Lakes Erie and Michigan, in 1983 for Lake Superior, and in 1985 for Lake Ontario. Great care was taken during collection, storage, and analysis to prevent sample contamination and to document any contamination occurring. Trace elements measured by atomic absorption techniques were silver, aluminum, arsenic, boron, barium, beryllium, bismuth, cadmium, cobalt, chromium, copper, iron, mercury, lithium, manganese, molybdenum, nickel, lead, antimony, selenium, tin, strontium, vanadium, and zinc. All results were field and laboratory blank corrected. Excluding aluminum, barium, iron, and strontium, concentrations of trace elements in most of the Great Lakes were a few ppb or less, with many elements being below one ppb. Element concentrations were highest in Lakes Erie and Michigan and lowest in Lakes Huron and Superior. All five Great Lakes had more than 50% of their total iron, aluminum, and manganese associated with paniculate matter.

Spatial patterns and temporal trends in mercury concentrations, precipitation depths, and mercury wet deposition in the North American Great Lakes region, 2002-2008

Annual and weekly mercury (Hg) concentrations, precipitation depths, and Hg wet deposition in the Great Lakes region were analyzed by using data from 5 monitoring networks in the USA and Canada for a 2002-2008 study period. High-resolution maps of calculated annual data, 7-year mean data, and net interannual change for the study period were prepared to assess spatial patterns. Areas with 7-year mean annual Hg concentrations higher than the 12 ng per liter water-quality criterion were mapped in 4 states. Temporal trends in measured weekly data were determined statistically. Monitoring sites with significant 7-year trends in weekly Hg wet deposition were spatially separated and were not sites with trends in weekly Hg concentration. During 2002-2008, Hg wet deposition was found to be unchanged in the Great Lakes region and its subregions. Any small decreases in Hg concentration apparently were offset by increases in precipitation.

Atmospheric transport of speciated mercury across southern Lake Michigan: Influence from emission sources in the Chicago/Gary urban area.

Quantifying the local and regional impacts of speciated mercury (Hg) emissions from major urban and industrial areas is critical for understanding Hg transport and cycling in the environment. The Chicago/Gary urban area is one location where Hg emissions from industrial sources are significant and the regional transport of emissions may contribute to elevated ambient Hg concentrations at downwind locations. From July to November 2007, we collected semi-continuous measurements of gaseous elemental Hg (Hg(0)), fine particulate bound Hg (Hgp), and divalent reactive gaseous Hg (RGM) in Chicago, IL and Holland, MI to characterize the impact of Chicago/Gary source emissions on Hg concentrations in southwest Michigan and to improve our overall understanding of speciated Hg transport and deposition. The mean (and median) concentrations of Hg(0), Hgp, and RGM in Chicago were 2.5ng/m(3) (1.9ng/m(3)), 9pg/m(3) (5pg/m(3)), and 17pg/m(3) (6pg/m(3)), respectively. In Holland the mean (and median) concentrations were 1.3ng/m(3) (1.3ng/m(3)), 6pg/m(3) (6pg/m(3)), and 8pg/m(3) (2pg/m(3)), respectively. Cluster analysis of 24-hour HYSPLIT back-trajectories associated with the semi-continuous Hg measurements indicated that southwest transport from Chicago/Gary to Holland occurred during approximately 27% of the study. In Holland, under this transport regime, we observed a five-fold increase in RGM relative to the median concentration of the other transport clusters. We applied the HYSPLIT dispersion model to two case study periods to further quantify the impact of Chicago/Gary sources on southeast Michigan and investigate the role of direct transport and dispersion of speciated Hg emissions. Results suggested that more than 50% of the maximum RGM concentrations observed in Holland during the selected periods could be attributed to direct transport of primary RGM emissions from Chicago/Gary. The remaining RGM fractions are believed to be associated with Hg(0) oxidation during transport over Lake Michigan.

Assessing the emission sources of atmospheric mercury in wet deposition across Illinois.

From August 4, 2007 to August 31, 2009, we collected event-based precipitation samples for mercury (Hg) and trace element analyses at four sites in Illinois (IL), USA. The objectives of these measurements were to quantify Hg wet deposition across the state, and to assess the contributions to Hg in precipitation from major local and regional emission sources. Monitoring sites were located, from north to south, in Chicago, Peoria, Nilwood, and Carbondale, IL. Measurements from these four sites demonstrated that a clear spatial gradient in Hg wet deposition was not evident across the state. Each site received>10μgm(-2) of Hg wet deposition annually, and these observed values were comparable to annual Hg wet deposition measurements from other event-based precipitation monitoring sites in source-impacted areas of the Midwestern U.S. We applied the multivariate statistical receptor model, Positive Matrix Factorization (EPA PMF v3.0), to the measured Hg and trace element wet deposition amounts at the four sites. Results suggested that 50% to 74% of total Hg wet deposition at each site could be attributed to coal combustion emissions. The other source signatures identified in the precipitation compositions included cement manufacturing, mixed metal smelting/waste incineration, iron-steel production, and a phosphorus source. We also applied a hybrid receptor model, Quantitative Transport Bias Analysis (QTBA), to the Hg wet deposition datasets to identify the major source regions associated with the measured values. The calculated QTBA probability fields suggested that transport from urban/industrial areas, such as Chicago/Gary, St. Louis, and the Ohio River Valley, resulted in some of the highest estimated event-based Hg wet deposition amounts at the four sites (potential mass transfer of up to 0.32μgm(-2)). The combined application of PMF and QTBA supported the hypothesis that local and regional coal combustion was the largest source of Hg wet deposition in Illinois.

Movement and Loadings of Inorganic Contaminants Through the Lower Saginaw River

Abstract The objective of this study was to investigate the movement of selected contaminants (four heavy metals) and conventional limnological variables through the lower Saginaw River. We related limnological variables to concentrations of contaminants, determined if sources of these contaminants existed in the lower 8 km of the river, and calculated loadings of materials to Saginaw Bay. Concentrations of most variables changed significantly among the six sampling dates in 1990–1991, while relatively few variables were different among the six sampling stations. Causes of changes in concentrations of variables in the river over time appeared to stem from two factors, seasonality and river discharge. One phase of the sampling program was conducted on six different dates which encompassed river flows ranging from greater than 700 m 3 sec −1 to less than 25 m 3 sec −1 . The date on which the least river discharge was observed included a flow reversal when Saginaw Bay water intruded into the lower Saginaw River. Statistical analyses showed that temporal variation among variables was due, in part, to fluctuating river discharge and in part to seasonal factors. The relative importance of these two components varied among parameters. For example, total suspended solids were positively correlated with river discharge, while chlorophyll varied more seasonally. In contrast to temporal trends, few parameters varied significantly among the six sampling stations. This was especially true for dissolved and paniculate metals. Loadings of key materials to Saginaw Bay from the Saginaw River were generally positively correlated with river discharge. We concluded that most substances entered the river upstream of our study section in the lower 8 km of the Saginaw River.

Mercury Distribution in the Biota of a Great Lakes Estuary: Old Woman Creek, Ohio

A study of mercury contamination in a wetland ecosystem was undertaken to assess degree of contamination in various compartments of the food web, and determine the most important routes for mercury to enter fish species. Old Woman Creek is a freshwater estuary on the southern shore of Lake Erie and has been recognized as an important nursery area for fish. Although there is no point source of mercury to the wetland, there is evidence of biomagnification of mercury in the food web. Mercury concentrations in water were not detectable (< 20 ng/L) by methods used in this study. The sediments had a mean concentration of 0.04 /μg total Hg/g wet weight. Zooplankton, benthos, and macro-phytes had less than 0.05 μg total Hg/g wet weight, but mean concentrations in fish tissue ranged from 0.001 to 0.636 μg total Hg/g wet weight. There was a positive correlation between mercury concentration and both length and weight in channel catfish Ictalurus punctatus and common carp Cyprinus carpio. In general, channel catfish accumulated higher levels of mercury than common carp, suggesting that the pelagic food web may be more efficient in delivery of mercury (to channel catfish) than the benthic pathway (to common carp). The highest value for fish occurred in a bowfin Amia calva. Tissue from one great blue heron Ardea herodias had the highest concentration at 1.62 μg total Hg/g wet weight. In this system, as in others, top predators tended to accumulate the highest mercury levels and were therefore most susceptible to mercury poisoning, even at sublethal levels.

Contamination of Green Bay water with lead and cadmium by a 37-m long, 2-m draft research vessel

During late April 1989, Green Bay, Lake Michigan water was sampled for dissolved concentrations of lead and cadmium. Samples used to document horizontal contamination of the water by a research vessel were collected using a rubber boat rowed 2, 50, 100, and 200 m upwind of the anchored mother ship. Samples used to document vertical contamination of the water column by the research vessel were collected from the vessel at water depths of 0.2, 2, 5, 10, and 20 m. Both lead and cadmium blanks were < 0.5 ng/l and below their limits of detection at the 95% level of confidence of 3.5 and 0.98 ng/l, respectively. Concentrations of lead in the horizontal direction varied between 3.5 ng/1 at 200 m from the ship and 7.7 ng/1 at 2 m from the ship. Cadmium concentrations varied from 2.8 ng/l at 2 m from the ship to 1.5 ng/1 at 200 m from the ship. Lead concentrations in the vertical direction varied between 8.4 ngll at a depth of 0.2 m and 3.3 ng/l at 5 m. Cadmium concentrations ranged between 4.5 ng/l at 2 m and 2.2 ng/1 at 20 m. The vertical studies were inconclusive and appeared to be influenced by resuspension of bottom sediments. Uncontaminated samples can be collected as close as 100 m to the research vessel.

An Artificial Turf-Based Surrogate Surface Collector for the Direct Measurement of Atmospheric Mercury Dry Deposition

This paper describes the development of a new artificial turf surrogate surface (ATSS) sampler for use in the measurement of mercury (Hg) dry deposition. In contrast to many existing surrogate surface designs, the ATSS utilizes a three-dimensional deposition surface that may more closely mimic the physical structure of many natural surfaces than traditional flat surrogate surface designs (water, filter, greased Mylar film). The ATSS has been designed to overcome several complicating factors that can impact the integrity of samples with other direct measurement approaches by providing a passive system which can be deployed for both short and extended periods of time (days to weeks), and is not contaminated by precipitation and/or invalidated by strong winds. Performance characteristics including collocated precision, in-field procedural and laboratory blanks were evaluated. The results of these performance evaluations included a mean collocated precision of 9%, low blanks (0.8 ng), high extraction efficiency (97%–103%), and a quantitative matrix spike recovery (100%).