Elemental Chemistry of Streambed Sediments of the St. Croix River Basin, 2000

U.S. Geological Survey Water-Resources Investigations Report 02-4087

By Mark E. Brigham

This report is also available as a pdf.

CONTENTS

Abstract

Introduction

Purpose and scope

Previous investigations

Environmental setting

Acknowledgements

Methods

Laboratory analysis

Elemental chemistry of streambed sediments

Implications for sources of trace elements

Namekagon River upstream and downstream of Hayward, Wisconsin

St. Croix River upstream and downstream of Stillwater, Minnesota-Hudson, Wisconsin

Kettle River upstream and downstream of Sandstone and Hinckley, Minnesota

Rush Creek upstream and downstream of Rush City, Minnesota

Osceola Creek upstream and downstream of Osceola, Wisconsin

Implications for water-resource quality

Summary

References

Appendix

ABSTRACT

Streambed sediments from 30 sites in the St. Croix River Basin were analyzed for selected chemical elements. Possible occurrences of low-level contamination include: (1) elevated concentrations of arsenic, lead, silver, and to a lesser extent cadmium, copper, and mercury in the Namekagon River downstream of Hayward, Wisconsin; (2) elevated lead concentrations in the St. Croix River downstream of the Stillwater, Minnesota—Hudson, Wisconsin region; (3) slightly elevated concentrations of Arsenic, Cadmium, Copper, Lead, Mercury, Silver, Vanadium, and Zinc in the Kettle River downstream of Sandstone and Hinckley, Minnesota; (4) and substantially elevated mercury concentrations in Rush Creek downstream of Rush City, Minnesota. Elevated concentrations of copper, lead, and to a lesser extent cadmium, in Osceola Creek, downstream of Osceola, Wisconsin may be due to anthropogenic sources or contributions from bedrock that differs geochemically from sediments farther upstream. There does not appear to be large anthropogenic contributions of trace elements—above the background geochemical and atmospherically deposited levels—in the streams sampled for this study. Many of the elements appear to be associated with mineral sources. Concentrations of potentially toxic trace elements in the St. Croix River Basin generally were lower than concentrations associated with frequent adverse effects to aquatic biota. Exceptions were arsenic (7 of 30 samples) and mercury (1 of 30 samples).

INTRODUCTION

Trace elements are natural constituents of the earth's crust; all bedrock and sedimentary deposits contain some levels of trace elements. Human activities such as mining, ore smelting, industrial processes, fossil-fuel combustion, and many other activities, may greatly augment trace element concentrations in the environment. Streambed sediments can be a useful medium for trace element analyses. Many trace elements strongly sorb to sediment particles. Streambed sediments can accumulate chemicals over time, and may be useful archives of past contamination. If significant trace element contamination is introduced to a stream—either transiently or continuously—streambed sediments should accumulate some portion of the elements through chemical and physical sorption processes. Processes that affect the fate of sediment-bound trace elements in rivers include sedimentation, burial of sediments in depositional areas (especially lakes and reservoirs), resuspension (especially during high flow events), desorption, and biological uptake.

Urban development in the St. Croix River Basin and the effects of potentially toxic contaminants and sediment from urban runoff have been identified as major issues in the basin by the St. Croix Basin Planning Team (Holmberg and others, 1997). Previous studies have found elevated concentrations of selected trace elements in the St. Croix River Basin. For example, fish-consumption advisories are established by Minnesota and Wisconsin because of potentially toxic levels of mercury in fish taken from the St. Croix River. A study by Troelstrup and Foley (1993) indicated that the higher concentrations of trace elements in mussel shells in the lower reaches of the St. Croix River Basin, compared to upstream reaches, corresponded to slower growth rates and sizes for those organisms. Troelstrup and Foley (1993) also compiled and analyzed existing water-quality and biological information from the Lower St. Croix National Scenic Riverway, finding concentrations of Cd, Cu, Fe, Mn, and Hg (element symbols are defined in the Methods section) in the water generally were greater in the lower reaches of the St. Croix River relative to the upstream reaches.

A better understanding of variability in trace element concentrations across the St. Croix River Basin, and links to anthropogenic sources could help water-resource managers understand threats to ecosystem health, which include the potential for toxicity to aquatic organisms, including endangered species. The U.S. Geological Survey (USGS), in cooperation with the National Park Service, undertook this study to better understand trace element levels and sources in the St. Croix River Basin.

PURPOSE AND SCOPE

The purpose of this report is to present the results of elemental chemical analyses of streambed sediment collected during June 19-August 3, 2000 at 30 stream sites in the St. Croix River Basin. Sites upstream and downstream of the larger cities in the basin were sampled to assess potential anthropogenic contributions of trace elements to the St. Croix Riverway. Concentrations of selected trace elements are compared to sediment-quality criteria and guidelines, providing an indication of potential toxicity concerns, which has implications for water-resource quality in the St. Croix River Basin.

PREVIOUS INVESTIGATIONS

Naturally occurring minerals in the St. Croix River Basin are one source of trace elements. Aluminosilicate minerals contain trace levels of many elements. Bedrock in the northeastern portion of the basin and in the Lake Superior Basin, where glacial sedimentary deposits in the St. Croix River Basin originate, is enriched with native copper (Nicholson and others, 1992).

In the absence of anthropogenic inputs of elements, it would be expected that the elemental composition of sediments reflects the composition of the parent geologic material (mineral phase), as modified by weathering and physicochemical reactions, such as dissolution, and precipitation. Some trace elements, such as mercury, typically have exceedingly low concentrations in the mineral phase (Smith, 1990); however, atmospheric transport and deposition from remote natural sources (volcanoes, and mercury volatilization from the oceans) have resulted in elevated concentrations of mercury in pre-industrial strata of lake sediments relative to local geologic sources (Brigham, 1992). The same processes probably contributed to elevated mercury concentrations in surficial soils and organic-enriched streambed sediments in the St. Croix River Basin, even prior to human settlement.

Humans have affected aquatic trace element cycling through direct discharges into streams and atmospheric emissions, which are deposited at varying distances from the source. Nriagu and Pacyna (1988) reported that human activities result in substantial inputs to the environment of Sb, As, Cd, Cr, Cu, Pb, Hg, Mn, Ni, Se, Sn, V, and Zn (element symbols defined in Methods section). Fossil fuel combustion, incineration of municipal and hospital wastes, metal smelting, and other emissions of metals have resulted in widespread releases of numerous elements into the atmosphere. Lead has been widely used as a gasoline additive, although it is no longer used in the United States. Industrial point-source inputs of trace elements to streams have been substantially reduced in recent decades in the United States (see, for example, Anderson and Perry, 1999). Acid mine drainage, and other mining-associated releases, continue to affect streams in certain localities, though with very few instances in the Upper Midwestern states. With reductions in direct discharges into streams, increased attention has focused on non-point sources, such as emissions to the atmosphere.

Proximity to anthropogenic sources has a substantially effect on trace element concentrations in streambed and lake sediments (Iskander and Keeny, 1974; Kroening and others, 2000; Rice, 1999; Verta and others, 1989). Particulate or highly particle-reactive forms of elements tend to be transported relatively short distances before they are deposited (Klein and Russel, 1973; Nriagu and others, 1998). In contrast, gaseous species, such as elemental mercury, can become widely distributed in the atmosphere (Fitzgerald and others, 1998).

Numerous studies point to the wide scale, combined loads of anthropogenic sources of trace elements to the aquatic environment. Sediment cores from lakes distant from point-source contamination have been used to assess changes in atmospheric deposition of trace metals over the past 150-200 years (pre-industrial to near present). Pronounced up-core (from preindustrial to recent sediment strata) increases in mercury concentrations have frequently been reported (Eng-strom and Swain, 1997; Meili, 1991; Swain and others, 1992; Verta and others, 1989). Other elements that often are enriched in more recent, near-surface lake sediments include Cd, Cu, Pb, and Zn (El-Daoushy, 1986; Iskander and Keeny, 1974; Kemp and others, 1978; Nriagu and others, 1979; Verta and others, 1989).

Kroening and others (2000) determined that streambed sediments in urban areas in the Upper Mississippi River Basin generally had higher levels of Sb, As, Cd, Cu, Pb, Hg, Ni, and Zn than non-urban areas. This reflects a national pattern of greater trace element concentrations (Cd, Hg, Cu, Pb, Zn, and to a lesser extent, Cr and Ni) in streambed sediments from urban compared to non-urban streams (Rice, 1999). Urban areas may include numerous sources of trace element emissions to the atmosphere, some portion of which fallout locally, the remainder being transported and deposited regionally or globally. Urban runoff and point discharges also may contribute trace elements to streams.

Kroening and others (2000) analyzed surficial streambed sediments and fish livers at four sites in the St. Croix River Basin. Results showed that concentrations of Cu, Pb, Hg, and Zn in streambed sediments were greater at the most downstream site sampled (located near Hudson, Wisconsin) relative to the other three sites sampled in less developed areas of the St. Croix River Basin.

ACKNOWLEDGEMENTS

Randy Ferrin and Barry Long (National Park Service), and Sharon Kroening (U.S. Geological Survey) were instrumental in initiating this study, and Randy Ferrin provided valuable input into preparation of this report. I also thank the following U.S. Geological Survey personnel: Luke Stuewe provided careful assistance with sampling and sample processing, Laurel Woodruff and Tim Cowdery offered expertise on geology of the region; Barbara Scudder and Bernard (Bernie) Lenz offered helpful review suggestions.

ENVIRONMENTAL SETTING

The St. Croix River Basin drains 20,176 km² in Minnesota and Wisconsin, and includes greater than 15 major tributaries to the Namekagon and St. Croix Rivers (fig. 1). The Namekagon River and St. Croix River form the St. Croix Scenic Riverway, a National Scenic Riverway of the National Park Service.

Figure 1. Location of the St. Croix River Basin, surficial geology, ecoregions, the St. Croix National Scenic Riverway, and sampling sites.

Noncalcareous Superior Lobe glacial sediments overlie much of the St. Croix River Basin (fig. 1). Calcareous Des Moines Lobe glacial sediments of the Grantsburg sublobe overrode the Superior Lobe sediments in the western portion of the basin (Wright, 1972) (fig. 1), resulting in substantial mixing of Des Moines Lobe and Superior Lobe sediments.

The St. Croix Scenic Riverway is a relatively well-preserved and biologically diverse aquatic resource. More than 60 State and Federally listed endangered and threatened species have been found in the Riverway. Among the 40 species of freshwater mussels found, 15 are State listed and two are Federally listed as endangered. There are 10 species of fish that are State listed as threatened or endangered.

Although the water quality generally is considered to be good in the upper St. Croix and Namekagon Rivers, activities in the tributaries and their watersheds are affecting the overall health of the St. Croix Scenic Riverway (Troelstrup and others, 1993). Due to its proximity to the Minneapolis/St. Paul metropolitan area, there are continued increases in use and developmental pressure from population growth in adjacent counties.

METHODS

Streambed sediments were sampled at 30 sites in the St. Croix River Basin (fig. 1, table 1), following techniques of Shelton and Capel (1994). Samples were collected at 9 sites on the St. Croix River, 3 sites on the Namekagon River, and 18 sites on other tributaries. For tributary streams, typically two sites were sampled—one upstream and one downstream of a city. Samples were collected during June 19-August 3, 2000 during low-flow conditions.

At each wadeable site, samples were collected and composited from 5-10 depositional zones where fine-grained sediments accumulate during low flow. At each depositional zone, several samples of the upper 2-5 cm of streambed sediment were scooped by hand with a 5-cm diameter Teflon tube, and placed into a Teflon bowl for compositing. At deeper, nonwadeable sites, a large (23 x 23 cm) stainless steel Ekman dredge (Wild-life Supply Co.) was deployed from a canoe to retrieve streambed sediment. Surficial streambed sediments were sampled from the dredge using a Teflon tube, taking care to avoid sampling material that was within 1 cm of the walls of the dredge.

Each sample was thoroughly stirred. A portion of each sample was sieved using a new, acid-cleaned 63-µm (micrometer) Nylon mesh sieve (Gilson Co.) held in an acrylic sieve frame (Gilson Co.), using less than 250 mL (milliliters) of native water (stream water taken from the sampling site) to aid in sieving. After settling, supernatant water was decanted, and samples were transferred to clean paper containers suitable for soil and sediment trace element samples (Star Paper Tube, Inc., Salt Lake City, Utah), oven-dried at 50°C, and shipped for laboratory analysis. Both sieved and unsieved sediments were retained for archive purposes.

LABORATORY ANALYSIS

Samples were submitted to the USGS-Central Region Mineral Resource Team, then transferred to XRAL Laboratory Services (Don Mills, Ontario, Canada) for analysis. A list of analytes, and a summary of analytical procedures, is given in table 2. Further details of analytical procedures are given in the Appendix. The USGS-Central Region Mineral Resource Team routinely assesses quality-assurance procedures of contract laboratories (Murphy and others, 1997).

Reproducibility was assessed by submitting five sets of split samples for analysis. Of the elements that were detected at greater than 5 times the lower limit of detection, most of the replicate samples had absolute values of relative percent differences (RPDs) below 15 percent (table 2), the stated laboratory criterion for acceptable data (Appendix). Exceptions (RPDs exceeding 15 percent) are one sample pair each for As, Cd, Cu, Au, and Mo; two samples for Ce and carbonate carbon; and 3 samples for Cr and Ag. Ag was the only commonly detected element that yielded poor precision for all replicates (table 2).

ELEMENTAL CHEMISTRY OF STREAMBED SEDIMENTS

Aluminosilicate minerals are key components of fine (clay and silt-sized) inorganic sediments. These minerals are largely composed of lithophil elements (after Gold-schmidt's classification of the elements, Goldschmidt, 1954; Sposito, 1989). Many of the elements measured for this study are lithophils. Within an area of relatively uniform geologic source material, and in the absence of significant contamination, concentrations of lithophil elements would be expected to correlate with Al content (denoted by r[Al], where r is the Pearson correlation coefficient). Among the lithophils analyzed in this study, many correlate strongly (Ba, Ce, La, Li, Mg, Nb, K, Sc, Na, Sr, Th, and Ti; r[Al] > 0.65, p<0.0001) or modestly (Nd, Yb, and Y; 0.35< r[Al] <0.64, 0.0001<p<0.05) with Al content (fig. 2; table 3). The abundance of aluminosilicates or clay minerals may be controlling the concentrations of many lithophilic elements in streambed sediments in the St. Croix River Basin, particularly those elements with strong correlation to Al content.

The strong, positive correlations of Na and Sr with Al content in the St. Croix River Basin are in contrast to the Red River of the North Basin (RRB), for which Na showed no significant relation to Al, and Sr was negatively correlated with Al (Brigham and others, 1998). These two lithophil elements were considered indicative of evaporite minerals in the RRB, but are more likely to be associated with aluminosilicates in the St. Croix River Basin, due to their strong correlation with Al and other lithophilic elements.

Co, Ga, Ni, and Zn showed modest correlations with Al content in the St. Croix River Basin. These elements are chalcophils (ore loving) (Gold-schmidt, 1954), and tend to associate with sulfide mineral phases in comparison to lithophils, which associate with silicates. Co and Ni also are siderphils (iron loving) and lithophils. Depending on temperature and oxidation state during rock formation, these two elements may associate with sulfides, iron ores, or silicate materials. Furthermore, processes such as sulfide inclusion, isomorphic substitution, and adsorption to mineral surfaces may result in their association of chalcophil or siderophil elements with aluminosilicate minerals (Sposito, 1989).

Organic carbon dilutes Al content, as indicated by its strong negative correlation with Al (fig. 2; table 3). Se and Ag correlate negatively with Al content, which suggests that the source of Se and Ag is not aluminosilicates. Also, Se and Ag were the only two elements that correlated positively with organic carbon (r= 0.66 and 0.50, respectively), which suggests that these elements are enriched in surficial soils and sediments, possibly due to atmospheric deposition from industrial emissions. Anthropogenic emissions to the atmosphere of Se (Nriagu and Pacyna, 1988) and Ag (Agency for Toxic Substances and Disease Registry, 1990, chapter 5) are known to occur, which supports the hypothesis that concentrations of these elements are enhanced at the earth's surface, relative to deeper, organic-carbon-poor sediments. Given this hypothesis, it is surprising that other atmospherically transported elements, such as Pb and Hg, did not correlate significantly with organic carbon.

Cr and V (lithophils) and Fe (a siderophil; Goldschmidt, 1954) had no significant relation to Al content in St. Croix River Basin streambed sediments, but strongly correlated to Al content in RRB sediments (for V, this correlation was strong when samples from areas of the RRB with Cretaceous marine shales were excluded) (Brigham and others, 1998). One possible explanation for the lack of correlation of Cr, Fe, and V to Al content is that the St. Croix River Basin is over-lain by mixed glacial deposits from two distinct glacial lobes (calcareous Des Moines Lobe and siliceous Superior Lobe sediments), whereas the RRB is predominantly overlain by deposits from one glacial lobe (Des Moines Lobe).

Other elements that do not correlate (r[Al] <0.35) with Al content are As, Cd, Ca, Cu, Au, Pb, Mn, Hg, Mo, P, and carbonate carbon. Strong correlations between As and concentrations of Fe, Mn, and P were observed. The strong correlation of As (a siderophil/chalcophil element; Gold-schmidt, 1954) with Fe suggests a geologic source of As. Concentrations of As, Cd, Cu, Pb, and Hg may be slightly elevated in the St. Croix River Basin due to anthropogenic sources because of poor correlation with elements indicative of almuminosilicates.

IMPLICATIONS FOR SOURCES OF TRACE ELEMENTS

Many of the trace elements in streambed sediment appear to be closely associated with parent mineral sources, with possible low-level anthropogenic contributions of several elements at some sites. In comparison to streams draining highly urbanized parts of the Minneapolis-St. Paul metropolitan area (Kroening and others, 2000), and compared to other highly urbanized stream basins across the Nation (Callender and Rice, 2000; Rice, 1999), streams in the St. Croix River Basin generally had low levels of the trace elements that are commonly elevated from anthropogenic sources (Cd, Cr, Cu, Pb, Hg, Ni, and Zn). Emission of trace elements to the atmosphere, and subsequent transport and deposition, contribute low levels of several elements to relatively pristine areas (see Previous Investigations section).

In five cases, there were higher concentrations of selected trace elements at a site downstream of a city or urban area, compared to a site on the same stream, upstream of the urban area. These cases are discussed below (fig. 3; table 4). There does not appear to be large anthropogenic contributions of trace elements—above the background geochemical and atmospherically deposited levels—in the streams sampled for this study. There is no basinwide pattern of increased trace element concentrations downstream of urban areas. Possible causes of elevated trace element concentrations are speculative. Rigorous determination of specific sources was beyond the scope of this study.

Figure 3. Concentrations of selected elements in streambed sediments of the St. Croix River Basin, 2000. (Map number corresponds to Figure 1 and Table 1).