Environ. Sci. Technol. 2003, 37, 4694-4701
Modeling Precipitation and Sorption of Elements during Mixing of River Water and Porewater in the Coeur d’Alene River Basin LAURIE S. BALISTRIERI* U.S. Geological Survey, School of Oceanography, University of Washington, Box 355351, Seattle, Washington 98195 STEPHEN E. BOX U.S. Geological Survey, West 904 Riverside Avenue, Room 202, Spokane, Washington 99201 JENNIFER W. TONKIN Department of Chemistry, University of Washington, Box 351700, Seattle, Washington 98195
Reddish brown flocs form along the edge of the Coeur d’Alene River when porewater drains into river water during the annual lowering of water level in the basin. The precipitates are efficient scavengers of dissolved elements and have characteristics that may make metals associated with them bioavailable. This work characterizes the geochemistry of the porewater and models the formation and composition of the flocs. Porewater is slightly acidic, has suboxic to anoxic characteristics, tends to have higher alkalinity, and contains elevated concentrations of many constituents relative to river water. Laboratory mixing experiments involving porewater and river water were done to produce the precipitates. Thermodynamic predictions using PHREEQC indicate that predicted amounts of ferrihydrite and gibbsite agree with removal of Fe and Al. Predictions of element removal by adsorption onto ferrihydrite are consistent with observed removal using a combination of surface complexation constants for the generalized twolayer model (As and Se), alternative surface constants derived from experiments at high sorbate-to-sorbent ratios (Cd, Co, Cu, Ni, Pb, and Zn), and adjusted surface constants to fit experimental data (Cr, Mo, and Sb). This new set of surface complexation constants needs further testing in other contaminated systems.
Introduction The oxidation of primary metal sulfide phases in ore deposits and mine wastes releases metals, sulfate, and acidity to the aquatic environment (1). The attenuation of acidity and concentrations of dissolved constituents away from the initial point of release depends on physical and biogeochemical processes including hydrologic transport, dilution with other sources of water, neutralization by acid buffering minerals, precipitation of minerals, and adsorption (2-7). Large ecosystems that have been impacted by historical mining activities provide natural laboratories in which to study processes that act to redistribute metals away from * Corresponding author phone: (206)543-8966; fax: (206)685-3351; e-mail:
[email protected]. 4694
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 20, 2003
their mineralized sources. An understanding of these processes is critical for successful management and remediation of these systems. One such ecosystem is the Coeur d’Alene River basin where a century of mining activities, hydrologic transport, and biogeochemical reactions have resulted in the dispersion of dissolved and particulate metals throughout the system (8-14). Our previous work in the Coeur d’Alene River basin focused on modeling metal removal from solution by precipitation and adsorption by oxyhydroxide and organic phases during the mixing of acidic groundwater or acidic adit drainage with near neutral surface waters (6, 15). The impetus for the present work is our observations of reddish brown flocs along the edge of the Coeur d’Alene River in the lower valley in late fall and their probable role in scavenging dissolved elements. The material is easily resuspended and transported downstream. The flocculent material adheres to plant surfaces after floodwaters recede and could be a potential source of metal uptake during feeding by waterfowl. We hypothesize that this flocculent phase is formed during the mixing of neutral Coeur d’Alene River water and slightly acidic, metal-enriched porewater that is draining from watersaturated, metal-enriched levee bank sediment during seasonal changes in water level in the basin. The goals of this work are to define the composition of porewater in water-saturated levee banks in the lower Coeur d’Alene River basin, model the formation and scavenging ability of the flocculent material, and evaluate our ability to describe element removal from solution in contaminated ecosystems using currently available thermodynamic data. Because the flocs are difficult to collect in their pure form due to incorporation of fine silt from the fluvial tailings deposits, we examine and model metal removal during the formation of flocs in laboratory mixing of porewater and river water.
Study Area The Coeur d’Alene River basin drains a large part of the north Idaho panhandle (see Figure 1). The swift flowing North and South Forks of the Coeur d’Alene River join to form the main stem of the Coeur d’Alene River. Because of a gradient change at Cataldo, water in the main stem of the river slowly moves through marshes and lateral lakes to Lake Coeur d’Alene. The outlet of the lake is the Spokane River. The Post Falls Dam, which is on the Spokane River, regulates the level of Lake Coeur d’Alene during a major portion of the year. The lake level is held at 647.7 m above sea level during the summer. At this elevation, the lake water backs up into the Coeur d’Alene River resulting in little, if any, current between Cataldo Flats and Lake Coeur d’Alene. Draw down of the lake and the back flooded river system occurs through regulation by the dam during the fall. For example, the level of the lake was lowered about 2.2 m during the fall of 2000. The level of the lake is not regulated during the rainy, winter season. The South Fork of the Coeur d’Alene River and its major tributaries pass through the world class Coeur d’Alene mining district (16) (see Figure 1). Ore deposits are steeply dipping quartz (SiO2) and siderite (FeCO3) veins containing stratigraphically controlled Pb-Zn-Ag ore shoots in Precambrian rocks of the Belt Supergroup (17-23). Veins are separated into two major types by ore mineralogy: (1) lead- and zincrich veins having argentiferous galena (PbS) and sphalerite (ZnS) and (2) silver-rich veins having argentiferous tetrahedrite ([(Cu, Ag)10(Fe, Zn)2Sb4S13]) and minor galena and sphalerite. Pyrite (FeS2), chalcopyrite (CuFeS2), and pyrrhotite 10.1021/es0303283 CCC: $25.00
2003 American Chemical Society Published on Web 09/17/2003
FIGURE 1. Map showing the Coeur d’Alene River basin in northern Idaho and sites at the edge of the river where river water and porewater were collected [Smelterville Flats (S1), Cataldo (C1), Rose Lake (R1), and Killarney (K1)]. (Fe1-xS) are locally abundant (24). Wall rocks around veins are altered and typically contain 10-15% carbonate minerals, including siderite (FeCO3), ankerite [CaFe(CO3)2], and calcite (CaCO3) (25). Milling and ore concentration practices varied over time in the district because of changing technology and economics. Early ore separation methods (prior to 1915), which included coarse crushing and gravity (or jig) mineral separation methods, were not very efficient and resulted in metalenriched tailings (10). Development of more efficient flotation methods between 1915 and 1925 resulted in tailings with finer grain size and lower metal contents. Approximately 56 million metric tons of metal-enriched tailings were dumped into the Coeur d’Alene River and its tributaries before environmental regulations required the installation of tailings ponds in 1968 (26). Because flow is faster in the upper Coeur d’Alene River system, the major repositories of discharged mine tailings are in the channel and flood plain of the lower Coeur d’Alene River between Cataldo and Harrison and in Lake Coeur d’Alene (11, 12, 27, 28).
Methods A reconnaissance study of porewater and associated sediment composition was conducted during the week of November 2-9, 1998. Detailed descriptions of these sites and results are presented elsewhere (29). Based on results from the 1998 study, river water near Cataldo and porewater samples from Smelterville Flats in the upper valley (S1) and from three previously occupied sites (C1, R1, and K1) in the lower valley (see Figure 1) were collected during November 6-8, 2000 for the mixing experiments. We specifically chose the sampling time to correspond with the annual lowering of lake level and the presence of flocculent material along the edges of the river. Sample Collection and Handling. Porewater was collected in plastic syringes using a sipper array that was inserted into the sediment at the edge of the river (30). Porewater was collected at discrete depths that ranged from 10 to 25 cm. After sample collection, a three-way valve on each syringe was closed to the atmosphere to minimize oxidation of the sample. About 4 L of Coeur d’Alene River water were collected into prerinsed 1-L plastic bottles near Cataldo using a peristaltic pump and an in-line 0.45 µm filter capsule. Upon return to the field-based laboratory, the pH of the samples was measured. Subsamples for metals, Fe(II), anions,
alkalinity, and dissolved organic C (DOC) were filtered through 0.45 µm nylon filters into acid-cleaned plastic bottles (Fe(II) and metals) or into prerinsed plastic bottles (anions, alkalinity, and DOC). Samples for metal analyses were preserved by adding 1 drop of redistilled, concentrated nitric acid per 30 mL of solution. Concentrations of Fe(II) were determined colorimetrically using 1,10-phenanthroline (31), and alkalinity was determined by Gran titration (32) in the field. Anion samples were kept cold, and DOC samples were frozen after returning to the main laboratory. Mixing Experiments. At the field-based laboratory, porewater from discrete depths at each site was filtered through 0.45 µm nylon filters and combined to provide enough porewater for the mixing experiments. This combined, filtered porewater sample is called composite porewater throughout the remainder of the paper. Subsamples were immediately taken from the filtered composite porewater and filtered river water for metals, Fe(II), anions, alkalinity, and dissolved organic carbon (DOC). These samples were used to obtain initial concentrations of constituents in composite porewater and river water (i.e., end member solutions). The remaining filtered river water then was spiked with 2 mL of 0.28 M KCl per 1 L river water to provide a conservative tracer for the mixing experiments. The composite porewater from each site and the spiked river water were mixed in ratios of approximately 0:10, 2:8, 4:6, 6:4, 8:2, and 10:0 (composite porewater: river water) using a graduated cylinder and placed into preweighed 250 mL acid-cleaned plastic bottles. The total volume of each mixture was 70 mL. These solutions were returned to the main laboratory, weighed, mixed periodically, and equilibrated for 8 days to allow for complete oxidation of the sample. After equilibration, pH was determined, and subsamples for metals, Fe(II), anions, alkalinity, and DOC were filtered and preserved in the same manner as the porewater samples. Concentrations of Fe(II) were found to be e3% of the initial Fe(II), confirming oxidation of the samples. The newly formed precipitates were collected by settling and decanting the supernatant and stored in 5 mL plastic bottles. Analytical Methods. Concentrations of dissolved metals in porewater, river water, water from the mixing experiments, and standard reference materials (SRM) (USGS T-135, USGS T-137, NIST 1643d) were determined by inductively coupled plasma-mass spectrometry (ICP-MS) (33). Good agreement between dissolved concentrations of Fe(II) determined by VOL. 37, NO. 20, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4695
TABLE 1. Composition of Filtered River Water and Filtered Composite Porewater Collected in the Coeur d’Alene (CdA) River Basin in November 2000a CdA Riverb species
units
average (n ) 5)
pH Ca K Mg Na Cl SO4 alkalinity DOC Al As Ba Cd Co Cr Cu Fe Fe(II) Li Mn Mo Ni Pb Rb Sb Se Si Sr Zn
µM µM µM µM µM µM mequiv/L mg/L nM nM nM nM nM nM nM µM µM nM µM nM nM nM nM nM nM µM nM µM
7.21 250 540 150 100 510 200 0.54 0.7 85 6 210 21 3
