U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2004-5090
Vertical Distribution of Trace-Element Concentrations and Occurrence of Metallurgical Slag Particles in Accumulated Bed Sediments of Lake Roosevelt, Washington, September 2002
Data Collection and Analysis
Sediment cores were collected from near the pre-reservoir river channel where the accumulation of reservoir sediment was thought to be thickest and least likely to be disturbed by the seasonal fluctuation in water level in the reservoir. Other considerations for site selection included avoiding the presence of large landslides along the margins of the reservoir and consultations with tribal archaeologists to avoid possible disturbances of submerged ancestral tribal areas. Lacustrine sediments were not found in pre-reservoir channel at the most upstream site, consequently, the core for this area was collected away from the pre-reservoir channel toward the left bank of the reservoir on what is likely a submerged terrace.
Five sediment cores were collected from the lower and mid reaches of the reservoir and one sediment core was collected from the Spokane Arm of the reservoir. The locations of these sediment cores were determined using a Global Positioning System (GPS) to determine latitude and longitude (table 2) and are shown in figure 1. Core designations were assigned using the nearest river mile (RM). Cores collected along the pre-reservoir Columbia River channel include CCR-624 in the lower reach of the reservoir, and CCR-643, CCR-668, CCR-692, and CCR-705 in the mid-reach of the reservoir. Core CSA-8 was collected from the Spokane Arm of the reservoir. Sites of continuously accumulating sediments could not be identified in the upper and Northport reaches of the reservoir. The three most upstream cores were in areas where previously reported concentrations of trace elements typically were high. These sediment cores were selected to determine if the concentrations of trace elements varied within the accumulated sediments of Lake Roosevelt. The other three sediment cores were collected farther downstream in the reservoir to determine if sediment concentrations might indicate the combined influence of inflow from both the Columbia and Spokane Rivers. These cores were collected from: (1) in the Spokane Arm of the reservoir outside of the general influence of inflow from the Columbia River (CSA-8), (2) in the mainstem of the reservoir upstream of the confluence with the Spokane Arm (CCR-643), and (3) downstream of the confluence of the Spokane Arm and the mainstem Columbia River (CCR-624).
Sediment samples with readily identifiable slag particles were collected at three sites, designated RSS-685, RSS-724, and RSS-743 (table 2 and fig. 1). Slag particles from these samples were used to characterize slag particles in Lake Roosevelt. Sediments in samples from the RSS-724 and RSS-743 sites were collected from beach deposits immediately adjacent to the reservoir and the Columbia River. Sediments in samples from the RSS-685 site were collected from beneath the reservoir sediments at a site where attempts to collect a sediment core to determine trace-element profiles was unsuccessful as less than 20 cm of lacustrine sediments were recovered. The slag containing sand underlying the lacustrine silts were interpreted to be a fluvial deposit of the pre-reservoir Columbia River.
Sediment cores used to develop trace-element profiles were collected using a Benthos gravity core with a diameter of 6.5 cm. The gravity core was suspended from a pontoon boat capable of collecting cores from depths as much as 120 m. The gravity cores were collected in September 2002. Because pore water could not be obtained from the gravity cores as originally planned, resampling of the three upstream sites was conducted in July 2003 using a box core with a 14×14-cm cross section. GPS coordinates were used to locate the original sampling sites.
Sample Collection and Processing
At each site where gravity cores were collected, several attempts were made to obtain undisturbed cores greater than 50 cm in length and that also encountered the pre-reservoir sediments. The least disturbed core with the greatest thickness of lacustrine sediments was selected for subsampling and analysis (fig. 1 and table 2). After collection, cores were subsampled by pushing the top of the core up through the plastic core liner and slicing the core into sections that were 2 to 5 cm thick. Sediment samples were packed on ice and shipped to the laboratory to be freeze-dried prior to submission for chemical analysis. Elemental concentrations were determined on concentrated-acid digests using a mixture of hydrochloric-nitric-perchloric-hydrofluoric acids and analyzed by inductively-coupled plasma/mass spectrometry (ICP/MS) or, in the case of mercury, cold-vapor atomic-absorption spectrometry. Total organic carbon in the cores was determined as the difference between total carbon and carbonate carbon. These laboratory analyses were conducted at the USGS laboratory in Denver, Colorado, using the procedures described by Arbogast (1996).
[Core and riverine sediment sample designation indicates nearest Columbia River mile as shown on 7.5-minute USGS quadrangle maps. Locations of sediment cores and riverine sediment samples are shown in figure 1. Latitude and longitude are given in degrees, minutes, and seconds. Abbreviations: cm, centimeter; m, meter; ft, foot]
| Core and riverine sediment sample designation | Length of core (cm) | Latitude(° ' ") | Longitude(° ' ") | Depth of water(m) | Elevation of upper surface of core(ft) | Length of subsection interval (cm) | Number of subsection intervals |
|---|---|---|---|---|---|---|---|
| Core along Impounded Columbia River | |||||||
| CCR-624 | 46 | 47 53 42 | 118 32 27 | 87.5 | 998 | 3 | 16 |
| CCR-643 | 57 | 47 57 18 | 118 21 07 | 74.7 | 1,040 | 3 | 19 |
| CCR-668 | 164 | 48 11 40 | 118 11 31 | 66.8 | 1,067 | 5 | 33 |
| CCR-692 | 38 | 48 30 26 | 118 10 51 | 45.7 | 1,136 | 2 | 19 |
| CCR-705 | 49 | 48 38 14 | 118 05 09 | 19.5 | 1,222 | 2 | 19 |
| Core along Impounded Spokane River | |||||||
| CSA-8 | 90 | 45 56 02 | 118 12 03 | 34.4 | 1,172 | 3 | 28 |
| Riverine Sediment with Slag | |||||||
| RSS-685 | - | 48 24 39 | 118 12 14 | 165 | - | - | - |
| RSS-724 | - | 48 49 12 | 118 56 08 | 0 | - | - | - |
| RSS-743 | - | 48 58 15 | 117 38 50 | 0 | - | - | - |
Quality-assurance and control procedures were incorporated into the study to assure that the resulting data were of a known and acceptable quality. Potential for contamination was minimized by using clean field procedures including dedicated core liners for each core. Equipment used for subsampling the cores into individual sections was thoroughly cleaned between each section using a non-phosphate soap and water. After washing, the equipment was thoroughly rinsed first with tap water followed by additional rinsing with deionized water. A variety of quality-control samples, generated in the laboratory and the field, were used including blanks, and replicate and standard reference materials. For each core submitted to the laboratory as a series of individual sections, an environmental replicate and the standard reference sample also were submitted as blind-quality assurance samples. The standard reference materials used as blind QA samples submitted from the field included the National Institute of Standards and Technology's Buffalo River Sediment number 2704 and National Research Council of Canada's PACS-2. The alert criteria for the analysis of the standard reference material was 80 to 120 percent of the certified concentration.
Pore-water and sediment samples were collected from the three upstream sites (CCR-668, CCR-692, and CCR-705) using a 14-cm box core. The box core was subsampled at three horizons; the surface to 1 cm depth, and the intervals from 1 to 2 cm, and 9 to 11 cm. Core subsampling was done immediately after the sampler was returned to the boat deck minimizing exposure to the atmosphere. The sectioned samples were placed in polyethylene bags, which were flushed and purged with pure nitrogen gases to remove oxygen before sealing then placed in a second polyethylene bag that also was purged with nitrogen before sealing. Exposure time to the atmosphere typically was less than 5 minutes. Samples were returned to mobile laboratory and transferred to 100 mL polyethylene centrifuge tubes. The headspace in the centrifuge tubes was purged with nitrogen and sealed with polyethylene film and spun in a centrifuge for 30 minutes. The pore water was decanted from the centrifuge tube, filtered using 0.45 micron filter, and acidified with ultra-pure nitric acid. The sediment was frozen with dry ice and a portion was sent to the laboratory for analysis along with the pore-water samples. Pore-water samples were analyzed for trace-element concentrations by ICP/MS. Corresponding sediment samples were analyzed for trace elements as described above for the determination of trace-element profiles. Splits of four of the sediment samples, two each from CCR-668 and CCR-705, were treated for 30 minutes at 50oC with 0.25 M solution of hydroxylamine hydrochloride to remove trace elements sorbed to the sediment surface and in iron and manganese hydroxide coatings on sediment surfaces (Chao and Zhou, 1983). The residual sediment after leaching was then analyzed for trace elements in the same manner as the untreated sediment samples.
Cs-137 concentrations were measured by gamma counting and reported as picoCuries per gram (pCi/g). In the laboratory, sediment samples for elemental analysis were weighed, frozen, freeze-dried, weighed again, and then ground to a fine powder. Wet and dry weights were determined and used along with an assumed density of solids of 2.5 g/cm3, to estimate porosity and bulk density used to calculate mass accumulation rates in the cores. As a check on assumed densities, measured densities were determined for 20 percent of samples from volumes of homogenized wet sediment, which were weighed, dried, and then weighed again so that bulk density and the density of solids could be calculated. Measured densities were about 2.2 to 2.5 g/cm3 and were not sufficiently different from the assumed density to warrant individual analysis.
Isolation and Identification of Slag Particles
Samples from selected intervals of the cores used to determine trace-element profiles were examined for slag-like particles. Intervals were selected based on elevated concentrations of trace elements identified in the analysis of slag samples, which included zinc, copper, antimony, and silver. Chemical and physical characteristics of relatively unweathered slag particles were determined from slag particles isolated in sample RSS-743 collected from a sand bar in the Columbia River (RM 743), about 1.5 km downstream of the International Boundary. Slag particles were more weathered in samples from RSS-724 and RSS-685 sites. Slag was characterized using a petrographic microscope and a Noran 660 SSI scanning-electron microscope equipped with an energy-dispersive X-ray spectrometer (SEM/EDS). Heavy mineral fragments and slag particles were separated from selected core sediments in which slag particles were not readily observable with a binocular microscope. The sediments were suspended in sodium-polytungstate solution (density of 2.9 g/cm3) and then the heavy fraction was separated by centrifugation (Commeau and others, 1992). Slag particles from the heavy fractions were selected for examination by petrographic microscope and semi-micro elemental analysis using the SEM/EDS.
For more information about USGS activities in Washington, visit the USGS Washington District home page .
