Validating Metal(loid) Flux Predictions from Lake Coeur d'Alene Sediments Using Contaminated Ponds as MesocosmsEPA Grant Number: F5B20286
Title: Validating Metal(loid) Flux Predictions from Lake Coeur d'Alene Sediments Using Contaminated Ponds as Mesocosms
Investigators: Toevs, Gordon R.
Institution: University of Idaho
EPA Project Officer: Lee, Sonja
Project Period: August 1, 2005 through August 1, 2006
Project Amount: $105,000
RFA: STAR Graduate Fellowships (2005) RFA Text | Recipients Lists
Research Category: Academic Fellowships
Lake Coeur d’Alene (CDA) in Idaho is the second largest natural lake in the Inland Northwest and lies within the boundaries of the expanded EPA COEUR D'ALENE BASIN SUPERFUND SITE. Lake CDA provides drinking water for area residents and serves as a primary recreational area for inhabitants of the Pacific Northwest. Over the last century Lake CDA became, and continues to be, the major collecting bed for contaminated sediments produced from over a century of mining and ore processing activities along the South Fork of the CDA River. As a result of these mining activities tailings enriched in As, Cd, Pb, Zn, and other trace elements were deposited in stream banks and bars along the South Fork and main stem of the CDA River. These materials have been regularly resuspended during periods of high water flow and secondarily transported into Lake CDA. The USGS has estimated that as much as 85% of the lake bottom is contaminated with metal(loids) and that the Lake may contain 75 million metric tons of contaminated material .
The overriding concern of management agencies responsible for lake water quality is the potential release of the accumulated metal(loids) into the overlying water column. However inadequate information exists at this time to make accurate predictions of metal(loid) release. Previous research has confirmed the possibility of release as metal(loid) concentrations in the sediment porewater can be many-fold higher than the overlying water column , the sediments support microbial communities in which Fe(III) and sulfate reducers are abundant , and microcosm enrichment studies demonstrated transitory contaminant solubilization [4, 5]. Previous research has led to the prediction that eutrophication will promote the release of trace elements . The Coeur d'Alene Tribe, EPA, Idaho Department of Environmental Quality, and local citizens groups need to know phase associations, contaminant stability, and the environmental conditions that would promote release in order to develop a plan for managing the ever-increasing use of this lake resource. To date, projections of metal(loid) release are based on models which may not accurately describe geochemical principles controlling the important processes. I will define these processes so accurate information can be supplied for modeling contaminant release.
Our objective is to determine how changes in the trophic status of Lake Coeur d’Alene, Idaho will affect the flux of metal(loids) from contaminated lake sediments and their potential risk to contaminate the overlying water column classified as a primary contact, domestic water supply resource by the state of Idaho. The sediments of Lake Coeur d’Alene are known to contain harmful levels of As, Cd, Pb, Zn, and other metals, and recreational areas frequently post warnings regarding local fish consumption and limiting contact events between children and the sediment. Currently there is insufficient data to accurately project metal(liod) flux as the biogeochemical processes controlling contaminant release have not been positively identified. Our null hypothesis is that changing trophic status will not have a significant impact on metal(loid) release and subsequently contaminate the overlying water column, negatively impacting human health and wildlife.
The relative importance of factors regulating the release of toxic metals from the sediment and their flux into the overlying water column remain controversial. In addition, prior research to characterize sediment geochemistry has produced contradictory results. In 1989 surface sediment grab samples collected in Lake Coeur d’Alene and analyzed using a two-step selective sequential extraction method determined the majority of the Pb, Cd, Zn, As, and Cu were associated with an operationally defined iron oxide phase . In 1990, twelve gravity cores were collected at various depths and in a variety of depositional environments . Numerous chemical and separation procedures were performed, but selective sequential extractions were abandoned because of concern that sample integrity had been compromised during handling and storage. However, other analyses and observations led to the conclusion that “most of the subsurface enriched trace elements are not associated with sulfide minerals; they probably are associated, like their surface sediment counterparts, with an operationally defined iron oxide phase” . In 1994, another group of investigators collected sediment cores from Lake CDA using methods designed to preserve redox status of the samples. Selective sequential extraction of these cores indicated the majority of the Pb, Zn, and As was associated with the sulfidic phase, a conclusion supported by redox measurements showing the highly reduced nature of subsurface sediments . Differing results between the research groups were further highlighted in a series of exchanges in which the extent of diagenesis and its role in sediment mineralogy and contaminant stability were debated [8-10].
Lacking the needed information, investigators have been forced to go forward with metal-flux measurements and modeling. Balistrieri  performed benthic-flux calculations and concluded that the sediment of Lake CDA appears to be a source of Zn, Cu, Mn, and possibly Pb to the overlying water. However, she acknowledged that the model input data available might have been inaccurate because oxidation occurred during storage and sample processing. More recent flux investigations were conducted with an in situ sampling device . The investigators concluded that the flux of metals from the sediment may be as significant to lake water as riverine inputs. However, this was a one-time sampling and the investigators recommended that pore water studies coordinated with sediment chemistry would provide critical information to interpret and develop models. Laboratory core incubation studies on sediment cores obtained from Lake CDA have been performed . Cores were incubated under anaerobic conditions in an attempt to measure benthic metal flux, but the results were inconclusive.
It is apparent one major issue that must be resolved prior to metal-flux modeling is the phase association of the contaminants as this is of primary importance to their stability [13, 14]. Our approach to resolving contaminant mineralogy and phase association is to characterize the sediments of Lake CDA from the surface to 36 cm in depth increments that correspond to the changing redox conditions of the sediments. The core sections have been preserved in a manner that minimizes any mineralogical or redox changes. The cores were sub-sampled and analyzed using the direct, non-destructive technique of X-ray Absorption Spectroscopy (XAS). XAS analyses are collaborative efforts with Dr. Scott Fendorf, Stanford University and Dr. Benjamin Bostick at Dartmouth College. This technique requires the selection of specific elements for analysis and very competitive time allocation at DOE synchrotron facilities. Through total elemental sediment digestions and pore water analyses we determined the three primary elements of interest for XAS studies were Fe, S and As. Additional solid-phase analyses completed to date include porosity, particle size determination, and mineralogical analysis using x-ray diffraction.
In order to establish relationships between the solid and aqueous phases, aqueous concentrations of ions present in the sediment porewater are needed. This is accomplished through the collection of sediment porewater in equilibrium dialyzers, modeled after other in situ sampling devices . The equilibrium dialyzers are inserted vertically into the sediments and retrieved after a 4-week equilibrium period. Upon retrieval, the samples are immediately removed from the dialyzers and preserved in accordance with acceptable procedures for the intended analytical technique. A portion of this porewater is then analyzed for major cations using Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) after acidification. Trace metal analyses is performed on acidified porewater samples by ICP-Mass Spectrometry (ICP-MS) and anion concentrations are determined from Ion Chromatography (IC) analysis on non-acidified, Fe-purged samples. The redox status of the sediments and the overlying water column are another important aspect of this study and are established from the profile of redox sensitive species. Elemental concentrations of the sediments are determined by analyses of hydrofluoric acid, Aqua regia microwave digest solutions analyzed by ICP-AES.
In addition to determining accurate geochemical characteristics for the lake, we propose that taking direct flux measurements from an equally contaminated environment that experiences annual redox cycles, can be used to validate the model. To address this need, we selected small, contaminated ponds serving as fish and waterfowl habitats in the lower Lake CDA Basin as mesocosms. Periodic fluctuations in the redox status of these ponds have been observed  and thus represent a unique opportunity to simulate redox reactions that might occur in Lake CDA. A site has been established in a wetland pond where we have been collecting seasonal samples in hopes of capturing the dynamics involved during seasonal changes. All of these samples have or will be analyzed in a similar manner as those of the lake. We are in the process of scheduling XAS analyses on the pond sediments.
Alkalinity of freshwater lakes is an important variable controlling toxicity of metals and is often used as a surrogate for a number of water quality characteristics that affect the toxicity of metals . Alkalinity is also an important component of establishing the neutrality of solutions with the use of a charge balance calculation. The challenge of this calculation in the CDA system is the large amount of Fe(II) in the porewater which immediately begins to precipitate when the equilibrium dialyzers are exposed to the atmosphere. As the likelihood of field titrations under anaerobic conditions is minimal, we have been developing a method that includes preserving total CO2 content of the samples, analyzing with a Gas Chromatograph-Mass Spectrometer (GC-MS), and modifying the basic equations for the carbon dioxide system  to determine total alkalinity. This work is on-going.
Two contaminated sites and a control site in Lake CDA were established and the data collection at these sites is complete. During each sample deployment, divers placed three equilibrium dialyzers at each location. When solid phase analyses were to be performed, the divers collected sediment cores at the time of dialyzer retrieval to minimize temporal and spatial variability. The contaminated sites confirmed the extensive metal(loid) contamination of the sediment as average concentration of As, Cd, Pb and Zn all exceeded the EPA levels for primary contact. However, the startling and very significant result was the magnitude at which the aqueous concentrations of some elements exceeded the EPA’s Aquatic Life Chronic Criteria (CCC). The average concentrations of As, Cd, Pb, and Zn of the porewater in μg L-1 were 455.4, 1.0, 24.4 and 196.8, respectively. These concentrations are remarkable when we consider the overlying water column, with the exception of Zn, typically meets Federal Water Quality Standards . Arsenic exceeds 1200 μg L-1 at numerous sampling depths indicating the dramatic potential and realistic concerns about lake water contamination through diffusion of these contaminants.
Arsenic in the overlying water column of Lake CDA was typically less that 5 μg L-1 so a physical or chemical barrier is preventing the diffusion of these high porewater concentrations. The presence of a physical barrier is unlikely as the core sub-sections we evaluated indicated the surficial sediment porosity exceeds 80%. As we examined the redox profiles from the interface into the sediments, we found the progression of terminal electron acceptors from NO3- to Mn(IV) to Fe(III) to SO42- was occurring very close to the interface. Identifying the exact depth of the sample-cell in relation to the sediment-water interface and thus the exact depth of these redox boundaries is difficult as this is a diffuse and variably stratified region . An additional challenge in this lake environment was the limited visibility and intermittently aggressive currents experienced during sampler placement and retrieval. However, the divers marked the approximate depth of the interface and there was frequently a sediment line on the dialyzers so we are confident that the redox transition is occurring within 10 cm of the interface. Additionally, NO3- was only present above the interface indicating an oxic cap. When this information was combined with the Fe and As profiles, it became apparent that iron oxides were forming at the interface and scavenging As from the porewater. As these oxides are buried through additional fluvial siltation, they transition to the suboxic zone, a zone of microbial iron reduction in which the Fe is reduced and any previously coprecipitated As is released to the porewater.
Fe-oxide barriers inhibiting trace metal diffusion into the overlying water column are well documented [21, 22], but the high levels of As in the porewater are unique. This was especially confusing as prior research had reported trace metal immobilization was occurring concurrently with SO42- reduction . We were recording high levels of As well below this zone. Additionally, we saw no impact on the aqueous Fe concentrations in the zone of SO42- reduction, leading us to believe that the high Fe:S ratio resulted in insufficient S to immobilize the metal(liods). To resolve this possibility we analyzed the sediment total elemental digest data and found the sediments averaged 1.7 mol of Fe kg-1 and 0.11 mol of S kg-1. This indicates there is only enough S to immobilize 0.11 mol of Fe as FeS and the formation of any other sulfidic mineral would be limited. Our next step was to determine the sulfur speciation in the solid phase so we could determine its significance in immobilization of metal(loids).
Sulfur X-ray Absorption Near Edge Structure (XANES) spectroscopy, one type of XAS, was performed at the Brookhaven Advanced Lightsource in Brookhaven, NY on beamline X-19A. XANES spectroscopy was used to speciate sulfur in Lake CDA sediments and identify specific components of the sediment solid phase that potentially participate in metal(loid) sorption. As mentioned previously, the presence of appreciable sulfur species would have positive implications for the stability of contaminants in this reduced environment. XANES is superior to extraction methods for speciation in that it is a nondestructive and direct technique that yields chemical and structural data concerning an element within the solid phase. XANES data clearly indicate that pyrite was the only S mineral above the level of detection and that its concentration increased with depth. Although a detrital origin of pyrite within the sediments seems possible if an oxic rind is formed during transport , its presence is more likely explained by diagenetic processes occurring within the sediments since galena or other sulfidic minerals indicative of mining activity were not detected. XANES data revealed the presence of inorganic SO42- in the upper 6 cm of the cores which confirms and corresponds to the suboxic region of the redox profile. Sulfur detected in the +6-oxidation state below 6 cm was organic, ester bound sulfate, which appears to become an additional source of sulfides at 24 cm, considerably below the anoxic zone at 12 cm. Further investigation is necessary to determine if organic, ester bound sulfate can be used as a terminal electron acceptor in these sediments. Additionally, sulfur with oxidation states consistent with that of thiols and sulfones were detected and showed relatively small changes in percentages of the total sulfur throughout the core sections. Our S speciation data indicates transitory phases consistent with diagenetic reactions and has implications for modeling sulfur’s influence on precipitation, dissolution, and stability of the contaminants.
Iron XANES and Extended X-ray Absorption Fine Edge Structure (EXAFS) spectroscopy analyses were conducted at the Stanford Synchrotron Radiation Laboratory (SSRL), Menlo Park, CA utilizing beamline 4-2 (8 pole wiggler) running under dedicated conditions. Preliminary analysis of these data indicate that Fe(III) is only present in the upper suboxic zone, supporting its presence at the interface to act as an As scavenger that prevents this contaminant from entering the overlying water column. There is also a trend of a metastable Fe(II) mineral like green rust carbonate forming in the suboxic zone and maturing to siderite with depth. One of the striking absences in these data was the presence of pyrite, which remained below the detection limit throughout the cores. Thus, although pyrite is present it plays only a negligible role in Fe(II) precipitation. Additionally, no iron mineral associated with metal(loid) contamination occurred above detection limits.
Arsenic XANES spectroscopy was conducted on beamlines 4-1 and 4-3 at SSRL. Preliminary results from these data confirm the presence As(V) and As(III) species at the surface, supporting the concept of As sorption to iron oxides. There is significant accumulation of As at the interface and a significant zone of depletion in the sub-oxic zone where the oxides would be undergoing dissolution and releasing the sorbed As. These zones of accumulation and depletion again support the concept of a clear Fe-As association. At the 12-cm zone of SO42- reduction, arsenopyrite formation occurs with formation also identified at the lower zone of proposed organic, ester-sulfate reduction. Orpiment (As2S3) and arsenopyrite accounted for the majority of the As in the anoxic zone. However, As(III) and As(V) were identified at all depths and all As phases tended to be transitory. This indicates the importance of understanding the biogeochemical relationships of As, as many of these transitions cause potential risk to human health and ecosystem recovery.
The next phase of my project is data collection and analyses from the pond site. We have completed 7 sampling sequences at this site, which include at least one event in each season. The sediment total elemental digests indicate that average contamination is greater than the lake and that maximum contaminant levels are substantially higher than the lake. The redox conditions are more compressed than the lake with the zone of SO42- reduction occurring within 5 cm of the interface. Additionally, NO3- has only been detected in the upper water column and Fe(II) is present at the interface indicating that this zone is suboxic and does not act as an oxic cap to inhibit As flux. One significant finding thus far is the remarkably low levels of As in the overlying water and the sediment porewater. There does not appear to be any particular trend and the majority of the concentrations are <5 g L-1 with the rare exception >10 g L-1. We are in the process of applying for XAS time at the appropriate synchrotron facility.
In summary, I have identified transitory phases of As in the sediments and porewaters of Lake CDA that pose a threat to human health and ecosystem recovery. We have identified the barrier that is preventing diffusion of As into the overlying water column as well as the processes that release As into the porewater. We have yet to identify the primary mineral that is responsible for As accumulation at the interface but my hope is that as we complete the As-XAS analyses of the wetland-pond sediments we will identify that source. At this juncture, when the lake and pond redox environments are compared, it appears the oxic cap in the lake is indeed preventing diffusion to the overlying water column. However, possibly more important is that it may also be responsible for the dissolution of the primary mineral transporting As to these environments, and therefore may be responsible for the acute toxicity that is currently present in the lake sediments. We have made significant progress toward our objective of identifying sediment geochemistry so models can be developed to evaluate changing trophic conditions. I will require one additional year to complete the present pond study, analyze the data, and publish our results. A one-year STAR Graduate Fellowship award would allow me complete these objectives.