Background levels of metals in St. Lawrence river sediments: Implications for sediment quality criteria and environmental management

Authors


Abstract

Preindustrial sediments dredged from the St. Lawrence Seaway are often considered potentially toxic relative to interim sediment quality criteria. The aim of the present study was to better document the background levels of target metals that were once used to define the minimal threshold of these criteria. Three extractions were performed on sediment samples to evaluate the distribution and potential bioavailability of trace metals. The results showed that the background levels established for the criteria are representative of contaminant concentrations observed in preindustrial sediments but not of the postglacial marine clays underlying these sediments. Chromium, nickel, and copper concentrations in postglacial marine clays, when solubilized by strong acids, exceeded the minimal effect threshold, whereas chromium and nickel frequently exceeded the toxic effect threshold and therefore are problematic in terms of applying sediment quality assessment criteria. The results suggest that these trace metals are mostly associated with inert silicates in postglacial marine clays and are unlikely to be bioavailable to aquatic organisms. Postglacial marine clays should be better considered differently than are sediments in regard to the sediment quality criteria implementation. Accordingly, the use of tools to identify the postglacial material may become necessary if a particular management disposition should be established for this sedimentary material. Total recoverable aluminum concentrations in sediments, in conjunction with other physical characteristics, was shown to be an interesting tool to identify this specific material. The normalization of metals concentrations with total recoverable aluminum concentrations would take account of natural mineralogical and textural variability of freshwater sediments in the quality assessment process.

INTRODUCTION

The interim criteria for assessing the quality of St. Lawrence River sediments (Canada) consist of 3 levels of quality assessment defined for targeted inorganic (arsenic [As], cadmium [Cd], chromium [Cr], copper [Cu], mercury [Hg], nickel [Ni], lead [Pb], and zinc [Zn]) and organic (polycyclic aromatic hydrocarbons, polychlorinated biphenyls, etc.) contaminants (SLC and MENVIQ 1992). The 1st level of assessment was named the no effect threshold (NET) and referred to background or natural concentrations of contaminants found in the available literature at the time (Loring 1978; Loring et al. 1978; Sérodes 1978; Barbeau et al. 1981; Gobeil et al. 1987; Pelletier et al. 1988, 1989; SNC-Procéan 1991; SLC and MENVIQ 1992). These characterizations involved very diverse sampling and analytical methods, and with the exception of the report by Sérodes (1978), all referred to the marine portion of the St. Lawrence and excluded postglacial marine clays.

The next 2 levels of quality assessment were determined by use of screening level concentration approaches and are referred to as the minimal effect threshold (MET) and the toxic effect threshold (TET), for which the contaminants were tolerated by 85% and 10%, respectively, of the benthic fauna studied (SLC and MENVIQ 1992). This approach was not based on observations relating sediment-associated metals toxicity to benthic organisms, using bioassays, but rather on the presence and absence of given benthic species at different metal concentration levels in sediment. These derived guidelines, therefore, do not directly account for metal bioavailability and background concentrations. The approach was also applied to the Great Lakes by the Ontario Ministry of the Environment (Persaud et al. 1992) and therefore allowed for the comparison of data banks and criteria established for the St. Lawrence River environment (SNC-Procéan 1991). When MET is exceeded, complementary investigations are required and sediment toxicity must be evaluated with the use of bioassays. On the other hand, when TET is exceeded, sediment must be confined or treated, without further investigations, and rehabilitation actions must be undertaken to rectify the situation.

Technical workshops were held in 1996 and 2000 to examine the interim criteria for assessing the quality of St. Lawrence River sediments. They found a number of problems related to the application of the criteria (PWGSC 1996; EC 2001), including the fact that local and regional background concentrations of several contaminants in the St. Lawrence River had yet to be assessed. Background concentrations of metals occurring naturally in sediments are information that is required during the implementation of toxicologically based sediment quality criteria. The information on background concentrations is considered when sediment quality criteria may be lower than the respective concentrations of naturally occurring metals (SLC and MENVIQ 1992; CCME 1995). Regional background concentrations are also useful in the development of site-specific sediment quality objectives (CCME 1995). Background concentrations could be established from reference sites (unaffected by point sources) or the preindustrial sediment horizon (Persaud et al. 1992). Postglacial marine clays, which are uncontaminated materials often dredged in the St. Lawrence River, represent an environmental management challenge as their metals concentrations can exceed background concentrations and even the quality criteria.

Table Table 1.. Composition of inorganic material in postglacial marine clays and St. Lawrence River sediments
Mineral phasesPostglacial marine clay (weight %)aSt. Lawrence River sediments (weight %)b
  1. a S. Lepage and L.-F. Richard, St. Lawrence Center, Montreal, PQ, Canada, unpublished data.

  2. b Loring 1976.

Quartz13.550–70
Feldspar24 (k-feldspar 6; plagio. 18)20–30
Mica17 (muscovite)<1
Clay minerals24 (chlorite 13; smectite 8; kaolinite 3)Trace
Amphiboles6,52–3
Pyroxene<1<1
Carbonates7
Oxides1.5 (magnetite, hematite, ± illmenite)1–3
Sulfides2.25 (pyrite ± pyrrhotite)Trace
Accessory mineralsTraceTrace
Organic carbon (present study: median by area)0.43–1.460.16–0.46

Preindustrial sediments dredged in the St. Lawrence Seaway were often considered potentially toxic relative to the interim sediment quality criteria and, thus, presented a concern for the management of the environment. As a result, these sediments required suitable containment at substantial cost (SLC and MENVIQ 1992). Interestingly, because the largest dredging operations in the St. Lawrence River predated both the establishment of the interim criteria and the adoption of environmental protection legislation, pre-industrial sediments dredged from the Seaway before 1992, often postglacial materials, were almost always deposited in open water, generally close to the dredging site (Villeneuve 2001).

Preindustrial sediments are generally defined as sediments that were deposited over 100 y ago, before the beginning of the industrial era. Sedimentation in the St. Lawrence River began mostly in the 19th century, after the completion of major waterway projects and construction of hydroelectric infrastructures (Villeneuve and Quilliam 2000; Pelletier and Lepage 2002). As a result, only a few areas of permanent sedimentation in the fluvial stretch of the St. Lawrence River, mostly in the fluvial lakes, present relatively thin layers of preindustrial records (Cremer 1979; Carignan et al. 1994). These sediments overlie the thick deposits of silty clays left behind by the Champlain Sea, which covered the entire St. Lawrence Valley during the last deglaciation period around 10,000 to 12,000 y ago (Cremer 1979). The latter clays are commonly called postglacial marine clays, although they often refer to both clays that were deposited within the Champlain Sea and clays from the Lampsilis lacustrine phase, which succeeded the Champlain Sea during the ice retreat period (J.-P. Guilbeault, BraQ-Stratigraphy, Montreal, PQ, Canada, personal communication). Postglacial marine clays are characterized by their invariable physical features, which include a typical bluish-gray color, a very fine grain size (between 95% and 100% silt and clay), and a very cohesive and compacted texture. Their organic carbon content varies between 0.1% and 2%, with median concentrations around 0.5%, whereas their moisture content ranges from 38% to 43%, with median levels of 41% (see Table 1). This sediment can be imagined as relatively unaltered rock flour, resulting from the erosive and destructive action of ice caps on Canadian Shield mountains (Cremer, 1979; Loring, 1991). Clearly, postglacial marine clays are considerably different from preindustrial and recent sediments deposited in the St. Lawrence River.

Because of these considerations and the problem related to the presence of elevated levels of contaminants in marine clays, a research project was launched in 1999 by the St. Lawrence Centre of Environment Canada to better document the local and regional background conditions for trace metals in St. Lawrence River sediments. A sediment sampling survey was conducted in both postglacial marine clays and the overlying preindustrial sediments in 3 fluvial lakes (Saint-Francois, Saint-Louis, and Saint-Pierre), whereas only postglacial marines clays were retrieved from the Vercheres area, which is subjected to strong, steady erosive conditions. The results of this project are presented in this article and complement existing information on preindustrial concentrations of contaminants, providing reliable data with which to establish background concentrations of targeted contaminants by use of the latest sampling and analytical techniques. The present study also appraises the application of guidelines for specific sediment materials through a geochemical sediment assessment. Sediment–metal associations, which were determined based on geochemical approaches, were used for comparisons with guideline-based sediment assessments. The information gathered under the project also contributes to the more judicious use of and eventual improvement to sediment quality assessment criteria.

Figure Figure 1..

Sampling station locations.

METHODS

Sediment sampling

Over a 3-y period beginning in 1999, a total of about 50 sediment cores and almost 200 samples were taken from the fluvial stretch of the St. Lawrence River. In the 1st year of the project (1999), sampling was conducted in the area of Vercheres and Contrecoeur, where 10 sediment cores and 30 samples were drawn in an area near the Seaway (Figure 1A). The 2nd year (2000) was devoted to the sampling of the northern portion of Lake Saint-Louis and the eastern side of Lake Saint-François, for a total of 20 cores and 100 sediment samples (Figure 1A). Finally, sampling activities in the 3rd and last year were devoted to the area between Sorel and Trois-Rivières, including Lake Saint-Pierre, for a total of 20 sediment cores and about 60 samples (Figure 1B). Sediment cores were collected by means of a percussion hammer corer. Preindustrial sediments were subsampled at different depths marking different deposition characteristics. Sediment sampling was performed as prescribed in the Methods Manual for Sediment Characterization (Environment Canada and MENV 1992).

Preindustrial sediment horizons in areas of permanent sediment accumulation were determined by use of data available in the literature such as lithological descriptions and sediment core geochronology (SNC-Procéan 1992; Carignan and Lorrain 2000). Sediment cores were subsampled for geochronological dating (i.e., Pb and cesium [Cs]) in areas where information was sparse or incomplete. Only sediments deposited before the industrial period were sampled from each sediment core. Immediately on sampling, subsampled sediments were frozen at −20 °C until analysis.

Sediment samples were analyzed for particle size, humidity, total organic carbon and nitrogen content, and trace metal concentrations. Samples were prepared and characterized for each analytical method according to the procedures recommended by Environment Canada and MENV (1992). Organic carbon and nitrogen were determined by use of a CHN analyser with a detection limit of 0.01% (Method 01–1090) (NLET 2002). The particle size distribution of sediment samples was evaluated after sieving by use of a laser sedigraph.

Metal analyses

Trace metal concentrations were analyzed by use of 3 different leaching procedures designed to define the distribution and potential bioavailability of trace metals within the sediment material. Extractions were performed on 3 distinct subsamples generated from each sediment sample. A 1st digestion, referred to as the total metal extraction, was performed with a mixture of perchloric, nitric, hydrochloric, and hydrofluoric acids, which dissolves all the phases and associated trace metals present in the sediments (Loring 1975; Loring and Rantala 1975; Gobeil 1987). A 2nd digestion, known as a total recoverable extraction, involved a combination of hydrochloric and nitric acid, also known as aqua regia, with the addition of perchloric acid, thereby solubilizing all the nondetrital phases in sediments and excluding most of the inert silicates (Lord 1982; Sturgeon et al. 1982; Canfield 1989; Krantzberg 1994).

Finally, a partial and less aggressive extraction was performed to solubilize the trace metals associated with the most reactive phases of the sediments, such as labile organic matter, iron (Fe) and manganese (Mn) oxyhydroxides, carbonates, and hydrous alumino-silicates (Huerta-Diaz and Morse 1990, 1992; Kostka and Luther 1994; Bono 1997). The latter extraction, a cold 1N hydrochloric acid leach, was chosen for its simplicity of execution, convenience for all trace metal analyses, availability of high-purity reagents, and recognition among the scientific community (Canfield 1989; Huerta-Diaz and Morse 1990; Leventhal and Taylor 1990; Canfield et al. 1992; Kostka and Luther 1994; Gagnon et al. 1995; Bono 1997; Lacey 1997; Saulnier and Mucci 2000). Total and total recoverable extractions were performed at the National Laboratory for Environmental Testing (NLET) of Environment Canada (Methods 02–2402 and 02–2403) (NLET 2002).

In addition, 1N hydrochloride (HCl)-extractable metals were recovered following the procedure described by Huerta-Diaz and Morse (1990) and analyzed by inductively coupled plasma–mass spectrometry (ICP/MS) and inductively coupled plasma–optical emission spectrometry (ICP/OES). The detection limits for the different parameters and extractions presented in this article are summarized in Table 2. Quality control for extractable metals met the requirements described in the Methods Manual for Sediment Characterization (Environment Canada and MENV 1992), whereas total and total recoverable metal extractions were analyzed as prescribed by the national Canadian Standards Association (CSA) Accreditation Program.

Table Table 2.. Detection limits
MetalTotal and total recoverable (mg/kg)Extractable (1N hydrochloride) (mg/kg)
Arsenic12
Cadmium0.10.02
Chromium11
Copper11
Mercury0.0040.01
Nickel0.51
Lead0.10.1
Zinc22
Aluminum5050
Iron5050
Lithium0.21
Magnesium5050
Manganese55

Extraction procedures and interim criteria

The different leaching procedures involved in the studies—from which the background levels of contaminants were established and which served as the basis for the interim criteria for assessing St. Lawrence sediment quality—included data obtained with the use of total extractions (Loring 1975, 1978; Loring and Bewers 1978; Sérodes 1978; Gobeil 1987) and other adapted chemical extractions such as concentrated nitric acid (Barbeau et al. 1981; Pelletier and Canuel 1988). The latter digestion is much less aggressive than are both total and total recoverable leaching (Snäll and Liljefors 2000).

By contrast, the data used by the Ontario Ministry of the Environment to establish MET and TET thresholds levels were essentially derived from metal concentrations determined by total recoverable leaching procedures (Persaud et al. 1992; Belles-Isles and Savard 2000). This methodological heterogeneity generated some uncertainty in the Quebec interim criteria document, as there was no indication of which type of extraction procedure should be used to assess sediment quality by threshold (SLC and MENVIQ 1992). Despite this apparent confusion, sediment characterizations in the St. Lawrence River are commonly carried out by use of total recoverable extraction, as recommended in the Methods Manual for Sediment Characterization (Environment Canada and MENV 1992). The interim criteria threshold values for the targeted trace metals are summarized in Table 3. As shown in this table, the NET and MET values for Cr, Cu, and Ni are identical. The documented background concentrations for all 3 metals were used to establish both of these threshold levels, as they exceeded the MET derived from the screening level concentration approach.

Table Table 3.. Interim criteria for quality assessment of St. Lawrence sediments (1992)a
Extractable trace metalLevel 1: no effect threshold (mg/kg)Level 2: minor effect threshold (mg/kg)Level 3: toxic effect threshold (mg/kg)
  1. a From Environment Canada and MENV (1992).

Arsenic3717
Cadmium0.20.93
Chromium5555100
Copper282886
Mercury0.050.21
Nickel353561
Lead2342170
Zinc100150540
Figure Figure 2..

Total trace metal concentration ranges.

RESULTS AND DISCUSSION

Trace metal concentrations

The Hg, As, Cd, and Pb concentration ranges and medians in postglacial marine clays and overlying preindustrial sediments are generally very similar for each leaching procedure, as illustrated in Figure 2a to c. This relationship highlights the low contamination level of the sediments sampled, as these elements are strongly related to anthropogenic inputs. Because total Hg is determined with an aqua regia digestion, in the absence of hydrofluoric (HF) acid, the results for total Hg are presented with total recoverable trace metals data (Figure 2b). In contrast, Cr, Cu, Ni, and Zn concentrations, for each chemical extraction, are consistently higher in postglacial marine clays than in preindustrial sediments (Figure 2).

The latter observation may originate most notably from the distinctiveness of the weathering processes involved in the formation of these marine clays, which could have led to an enrichment of mineral phases that are otherwise sparsely present in sediments (Loring 1991); this hypothesis is addressed in greater detail in subsequent sections of this article. Furthermore, postglacial marine clays are fine materials, and grain size can influence metal determination in sediments. Snäll and Liljefors (2000) reported that the efficiency of strong acid leaching procedures was greatly increased with decreasing grain size. This phenomenon may further explain the general enrichment observed in the postglacial marine clays compared with St. Lawrence River preindustrial sediments for almost all the elements analyzed in the present study. Moreover, because of their fine-grained composition, marine clays can also be subject to greater trace metal adsorption as a result of a higher site-specific adsorption density, which is well documented in the literature (Loring 1978; Lorrain et al. 1993).

Figure Figure 2b..

Total recoverable trace metal concentration ranges.

Figure Figure 2c..

Extractible trace metal concentration ranges.

A thorough examination of the data revealed that a significant proportion of the samples displayed higher total recoverable Zn and Cr concentrations than their respective total concentrations. Although statistically possible for metals known to be preferentially associated with non-refractory mineral phases, these analytical discrepancies defied the probabilities in the sediments sampled in 2000; this is especially true of Zn. In the case of Cr, these irregularities could be attributed to Cr precipitation with fluorine in the total digestion solution, thereby causing its concentration to be underestimated. Furthermore, even though the total digestion procedure is standardized with certified materials, it may not dissolve all the refractory mineral phases potentially present in sediments, including chromite, magnetite, and accessory minerals such as zircon and sphene (Actlabs, PSC Analytical Services, Inc., Montreal, PQ, Canada, personal communication). Nevertheless, the total recoverable Cr measured in the present study only exceeded total Cr by an average of 10%, which was considered acceptable.

However, total recoverable Zn concentrations in a strong majority of the samples retrieved from Lakes Saint-François and Saint-Louis exceeded their respective total concentrations by an average of 13% and 22%. The same observation could be made for lithium [Li] levels in both Lakes Saint-François and Saint-Louis, as well as Vercheres sediments, where the average deviation varied from 4% to 40%. This phenomenon might have been caused by matrix interference, which could have led to an overestimation of total recoverable Zn and Li, as suggested from the results of the certified materials analyzed simultaneously (National Laboratory for Environmental Testing, Burlington, ON, Canada, personal communication). When the results obtained for the different sites are compared, this explanation appears probable and may raise doubts with regard to the efficiency and precision of this type of chemical extraction for these elements.

Comparison to interim criteria

A comparison of the concentrations measured in both preindustrial sediments and postglacial marine clays against the interim criteria thresholds reveals interesting contrasts. To varying degrees, the concentrations in preindustrial sediments exceeded the 1st-level threshold (i.e., NET) for total Cd and Cu, both total and total recoverable As and Cr, and, to a lesser extent, Hg (Table 4). These amounts were predominant in sediment collected from Lake Saint-Louis followed by Lake Saint-François. Moreover, around one-third of these sediment samples showed total As and both total and total recoverable Cr concentrations beyond their respective MET levels. It must be noted again that the NET and MET are identical for Cr, Cu, and Ni. However, despite these observations, median and average total and total recoverable concentrations of all these elements, with the exception of total As and Cd, did not exceed their respective NETs for each area. The results of the 1N HCl extraction are not presented in Table 4 because this extraction method is much weaker than the digestion methods used to define the criteria thresholds.

Accordingly, the median and average levels of trace metals found in the preindustrial sediments analyzed in the present study correspond to those reported in the studies from which the NET were derived (Table 5). This similarity is stronger with the total extraction because, as mentioned previously, the data generated by those studies were mostly based on this extraction procedure and other similarly aggressive leaching procedures. Again, only total As and Cd presented median concentrations slightly above the natural levels reported by these investigators.

Although similar observations can be drawn from As and Cd concentrations in postglacial marine clays, a very different scenario was observed for other trace metals. Total and total recoverable Cr, Cu, and Ni in nearly all the clay materials taken in each sampling area exceeded the NET and MET, whereas varying proportions (10–100%) of total and total recoverable Zn exceeded the NET (Table 4). Furthermore, total and total recoverable Cr and Ni were also frequently above the TET. Amounts exceeding the TET can be as high as 95% in the clay material sampled in the Vercheres area and 78% in Lake Saint-François. This situation represents a challenge for environmental sediment management, as sediment exceeding the TET for one of the interim criteria–targeted parameters must be considered toxic to benthic fauna. Accordingly, these sediments must be treated or suitably contained, and the possibility of rehabilitating the so-called contaminated environment must be considered. This seems quite paradoxical when applied to the 10,000- to 12,000-y-old natural sediments that cover most of the bottom of the St. Lawrence River in near entirety, and which are particularly exposed along the length of the Seaway. Moreover, these clays comprise most of the river's suspended particulate matter. Nonetheless, Ni and Cr in their bioavailable forms are known to be highly toxic to plants and aquatic organisms.

Trace metal partitioning and potential bioavailability

The natural enrichment of postglacial marine clays in Cr, Ni, and Cu, most particularly, raises questions about their potential bioavailability and, thus, their distribution within mineral phases and operationally defined extractions. The relative proportion of total recoverable phases—as well as extractable phases, for Cu—are greater in postglacial marine clays than in preindustrial sediments, to the detriment of the HF-extractable portion (Figure 3). The same observation can also be made for a number of other elements including Pb, Cd, and major elements such as aluminum [Al], Fe, and magnesium [Mg]. This indicates that the general enrichment found in postglacial marine clays would be essentially attributed to the increased presence of minerals that are soluble in aqua regia and in 1N HCl, to a lesser extent, in comparison to preindustrial sediments in the St. Lawrence River.

Clay minerals and micas, essentially chlorite and muscovite, are predominant in postglacial marine clays, as opposed to quartz and feldspars in preindustrial sandy muds encountered in the St. Lawrence River (Loring 1975, 1976; S. Lepage and L.-F. Richard, St. Lawrence Center, Montreal, PQ, Canada, unpublished data) (Table 1). Whereas quartz composes up to 70% of St. Lawrence River sediments, it constitutes no more than 13% of postglacial marine clays. However, micas and clay minerals make up around 40% of postglacial marine clays, whereas only trace amounts of these mineral phases can be observed in St. Lawrence River sediments (Table 1). Postglacial sediments were also reported as sulfide poor (Table 1). Trace metals such as Cr, Ni, and Cu are unlikely to be associated with quartz and potassium-feldspar, but may be considerably enriched in clay minerals such as chlorite, and, more importantly, in sulfides and magnetite (Loring 1975, 1976). As such, their abundance in preindustrial sediments was far less likely than in postglacial marine clays (Table 1).

Table Table 4.. Frequencies of sediment samples exceeding the Interim criteria thresholda
Preindustrial sediments
MetalNETMETTET
%SFSLSPSFSLSPSFSLSP
Arsenic-T86681438290000
Arsenic-TR294870190000
Cadmium-T2910096000000
Cadmium-TR0610000000
Chromium-T2938729387000
Chromium-TR4332043320000
Copper-T29642964000
Copper-TR000000000
Mercury-TR0921007000
Nickel-T034034000
Nickel-Tr000000000
Lead-T0011000000
Lead-TR000000000
Zinc-T034000000
Zinc-TR034000000
No. samples73228      
Postglacial marine clays
MetalNETMETTET
%SFSLVrSPSFSLVrSPSFSLVrSP
  1. a NET = no effect threshold; MET = minor effect threshold; TET = toxic effect threshold; T = total; TR = total recoverable; SF = Lake St. François; SL = Lake St. Louis; Vr = Vercheres area; SP = Lake St. Pierre.

Arsenic-T50961577658500050
Arsenic-TR2262-62027-000-0
Cadmium-T2827259200000000
Cadmium-TR040000000000
Chromium-T1001001001001001001001007815950
Chromium-TR10010010085100100100857812850
Copper-T1001001001001001001001000000
Copper-TR100881003810088100380000
Mercury-TR005000000000
Nickel-T1001001001001001001001007812900
Nickel-Tr10010010085100100100857223150
Lead-T0410000000000
Lead-TR040000000000
Zinc-T7881807704000000
Zinc-TR100881046019000000
No. samples18262113        
Table Table 5.. Background concentrations in St. Lawrence river sediments
SourceAreaArsenic (mg/kg)Cadmium (mg/kg)Chromium (mg/kg)Copper (mg/kg)Mercury (mg/kg)Nickel (mg/kg)Lead (mg/kg)Zinc (mg/kg)
Carignan et al. (1994)Lake St. François0.156217321578
Pelletier and Lepage (2002)Lake St. François0.186312221855
Sloterdijk (1983)Lake St. François2.61.500.14
Barbeau et al. (1981)Saguenay200.101785
Loring (1975, 1976)Gulf of St. Lawrence53260.25262184
Gobeil et al. (1987)Laurentian Trough0.20
Present study: total recoverable metalsFluvial stretch3.30.1438130.03187.461
Present study: total metalsFluvial stretch50.3146200.03231560

Loring (1976) found that “detrital” Cr and Ni in Saguenay Fjord sediments were associated with ferromagnesian minerals such as chlorite, hornblende, biotite, and pyroxene, as well as oxide minerals. Loring, however, refered to the detrital fraction of metals as the portion that is soluble in a total extraction with the use of HF acid but insoluble in a mild digestion with 25% v/v of acetic acid. When compared against the analytical methods employed in the present study, this detrital fraction would include both the total and total recoverable trace metals. Loring also reported, from microprobe analysis, that magnetite may contain Cr concentrations as high as 100,000 mg/kg. Although sulfides may be present in small amounts, they may contribute significantly to trace metal concentrations in both extractable and total recoverable phases. The strong affinity of a number of trace metals for sulfides, specifically acid-volatile sulfide and pyrite, are well documented in the literature (Huerta-Diaz and Morse 1992; Morse and Arakaki 1993; Huerta-Diaz et al. 1998). Pyrite can contain up to 500 mg/kg of Ni and may contribute up to 25% of the total detrital Ni present in Saguenay Fjord sediments (Loring 1976).

Work by Snäll and Liljefors (2000), in leaching experiments conducted on selected minerals isolated in glacial tills derived from the Scandinavian Precambrian bedrock, demonstrated that biotite was completely soluble and chlorite almost entirely soluble in strong acids, such as aqua regia and even boiling 1N HCl. Microcline, plagioclase, hornblendem and muscovite were, however, only slightly soluble in aqua regia and other leaching solutions involving different concentrations of boiling HCl and nitric acids. According to the data generated by these investigators, digestion with boiling 1N HCl would also solubilize relatively similar amounts of major elements from different common silicates than aqua regia. This observation, when compared to the present study and those of others (Brumbaugh et al. 1994; Raiswell et al. 1994; Saulnier 1997), highlights the striking difference between the leaching capacities of boiling and cold HCl extractions. Trace metals recovered with the cold 1N HCl extraction performed in the present study only represented generally less than 50% of the portion solubilized with the total recoverable leaching procedure.

A cold 1N HCL extraction is reputed to solubilize loosely bound trace metals, also referred to as “exchangeable trace metals,” as well as trace metals associated with labile organic matter (Bono 1997), carbonates poorly crystallized oxides (Canfield 1989; Kostka and Luther 1994; Raiswell et al. 1994) and probably reactive sulfides such as acid volatile sulfides (FeS) (Huerta-Diaz and Morse, 1990). This extraction is also used for the determination of simultaneously extracted metals (SEM) associated with acid volatile sulfides (AVS) in wet sediments (Ankley et al. 1996; Berry et al. 1996; DeWitt et al. 1996; Hansen et al. 1996; Leonard et al. 1996; Sibley et al. 1996). Acid volatile sulfides are known to control significantly metal activity and toxicity in sediments (Di Toro et al. 1990).

Several studies highlighted the relation between sediment-associated metals toxicity to benthic organisms and SEM: AVS ratios and differences. Results of these studies suggested that the 1N HCl extraction most likely solubilizes the bioavailable fraction in sediments characterized by low acid volatile sulfides concentrations (Ankley et al. 1996; Berry et al. 1996; DeWitt et al. 1996; Leonard et al. 1996; Sibley et al. 1996). Although the cold 1N HCl digestion is operationally designed to solubilize sulfide reactive Fe phases (Canfield 1989) and associated trace metals, its efficiency in extracting metals associated with sulfide minerals in dried sediments was called into question (Raiswell et al. 1994; Mikac et al. 2000). This research raised doubts regarding the efficiency of this extraction method to solubilize oxidation-resistant sulfides (Brumbaugh and Arms 1996) and their oxidation products in dried sediments (Raiswell et al. 1994) and also stressed the possibility of trace metal sulfide precipitation within the leachate solution (Mikac et al. 2000). However, the results of more recent studies contradict the latter affirmations in the case of iron monosulphides, also referred to as acid volatile sulphides (A. Mucci, McGill University, Montreal, PQ, Canada, personal communication). However, Huerta-Diaz and Morse (1990) demonstrated that pyrite solubilization can only be achieved with the use of concentrated nitric acid and not a 1N HCl solution. Because metals may be remobilized among sediment fractions, the weak 1N HCl extraction may underestimate the potentially bioavailable trace metals over time. For this reason, this article also presents results for metal extractions using stronger acids.

Figure Figure 3..

Copper, chromium, and nickel distribution in sediments within the three leaching procedures, using median values. PGMC = postglacial marine clays; PI = preindustrial sediments.

Accordingly, the increased proportion of trace metals in the 1N HCl–extractable fraction in postglacial marine clays could be attributed to the presence of carbonates and possibly pyrrhotite, a potentially reactive sulfide. The absence of any relationship between extractable Fe and Cu and between Ni and Cr in postglacial marine clays (0.09 < r2 < 0.23) would confirm that carbonate-associated and adsorbed metals may represent an important fraction of the extractable trace metals in this material. In contrast, the fairly good relation observed between extractable Fe and trace metals in preindustrial sediments (0.69 < r2 < 0.82) tends to demonstrate that Fe oxides or oxyhydroxides and volatile sulphides could be an important source of associated trace metals (Figure 4).

Based on the literature and the data presented in Table 1, the total recoverable extraction (i.e., aqua regia) would have leached from postglacial marine clay samples, in addition to the fractions recovered by the 1N HCl digestion, trace metals associated with clay minerals such as chlorite; well-crystallized oxides, including magnetite and hematite (Loring 1976; Snäll and Liljefors 2000); refractory organic matter (Tessier et al. 1979; Bono 1997); and less reactive sulfides such as pyrite (Huerta-Diaz and Morse 1990; Huerta-Diaz and Morse 1992; Morse and Arakaki 1993; Brumbaugh and Arms 1996; Huerta-Diaz et al. 1998). In addition, this extraction may also have slightly solubilized the metals incorporated into amphiboles (≅20%) and muscovite (≅5%), feldspars (≅5%), and pyroxene based on the work of Snäll and Liljefors (2000). The fraction remaining with the extraction method would be the metals associated with inert silicate materials.

However, the trace metals extracted from St. Lawrence River preindustrial sediments by use of this leaching procedure may have mostly originated from refractory organic matter, well-crystallized Fe and Mn oxides, and possibly partial leaching of inert silicate minerals. Also, as discussed earlier, the feldspar and amphiboles present in preindustrial sediments may be less resistant to chemical attack resulting from more intense weathering processes, thereby releasing greater amounts of their trace metal content with less aggressive leachants.

The correlations observed between Fe, Mg, and Al with Cr, Cu, and Ni in all the sediments sampled in the present study (0.69 < r2 < 0.92) tend to confirm the importance of ferromagnesian silicates, most likely chlorite, to the distribution of these trace metals in both postglacial marine clays and preindustrial sediments (Figure 5a to c).

Figure Figure 4..

Relation between extractible iron and trace metals in preindustrial sediments from the St. Lawrence River.

Lastly, the additional trace metal concentrations leached by the total digestion with the use of HF would be attributed, for both postglacial marine clays and preindustrial sediments, with the dissolution of the most inert silicates, including quartz, feldspars, amphiboles, and pyroxene (Loring and Rantala 1975; Snall and Liljefors 2000). With the exception of major elements such as Al, the latter phases would contribute, to a minor extent, to most trace metal concentrations in sediments owing to their low affinity toward these elements (such as quartz) or their relatively low abundance (such as pyroxene and amphibole). The results shown in Figure 3 emphasize the impoverishment of postglacial marine clays with respect to these mineral phases, essentially quartz (Table 1), in comparison to St. Lawrence River preindustrial sediments.

To summarize, roughly 20% of the Cr and 35% of the Ni present in postglacial marine clays, as well as in St. Lawrence River preindustrial sediments, would be associated with the commonly defined reactive phase and thus potentially bioavailable. However, a large proportion of these metals (˜60–75%) would be associated with more inert phases, mainly ferromagnesian silicates. Only 6% of these metals would be incorporated into the detrital fraction of postglacial marine clays, whereas this proportion would rise to 20% for St. Lawrence River preindustrial sediments.

Normalization approach

Despite their notable differences, the distribution of trace metals and major elements follow a continuum in both types of sediment. The linear regressions displayed in Figure 5 express the relationship between different elements in natural uncontaminated sediments. The background levels of contaminants, as determined in natural sediments, are frequently normalized and used to evaluate anthropogenic inputs in recent sediments (Loring 1991; Aloupi and Angelidis 2001). Background levels of contaminants play a key role in sediment quality assessment and are important to restoration projects. Therefore, contaminant concentrations are often standardized by use of parameters such as clay and organic carbon content, as well as conservative elements such as Al and Li, to compensate for the mineralogical and textural variability of sediment (Ackerman 1980; Loring 1991; Aloupi and Angelidis 2001). Although normalizations to a reference element are often performed for many metals in marine sediments, its applicability remained to be investigated for metals in freshwater sediments.

When looking at the results of this study, no strong relationships emerge between trace metal concentrations and clay or organic matter content. However, there was good correlation between Cu, Ni, and Cr and Li, as well as Al, concentrations in all the sediments sampled for the present study (Figures 5 and 6). These correlations were also confirmed with other trace metals such as Pb and Zn, and major elements such as Fe and Mg. In the case of Li, the linear relationship was much stronger when total element concentrations (0.85 < r2 < 0.96) were used than were those from the total recoverable fraction (0.54 < r2 < 0.82). This observation supports the use of Li to normalize St. Lawrence River sediment by use of total metal concentrations, as presented by Loring (1991). Nevertheless, because Li is principally incorporated in micas and ferromagnesian clay minerals (Loring 1991), which are almost entirely dissolved by aqua regia (Snäll and Liljefors 2000), total recoverable Li would have been expected to correlate well with extractable trace metals. The weaker correlation observed with total recoverable Li and trace metals in the present study probably reflects the analytical anomalies observed with total recoverable Li, as described previously. However, total recoverable Al displayed very good correlations with Cu, Ni, and Cr as well as Zn, Pb, Fe, and Mg (0.70 < r2 < 0.92). Not surprisingly, total Al did not relate to the latter elements because it is a major constituent of most mineral phases present in sediments.

Figure Figure 5..

Relation between total recoverable iron (a), magnesium (b), and aluminum (c) with total recoverable trace metals from sampled sediments. PGMC = postglacial marine clays; PI = preindustrial sediments.

As discussed above, total recoverable digestion only placed in solution aluminosilicates with strong affinities for trace metals and, more importantly, Cr, Ni, and Cu, whereas the total extraction also dissolves inert Al-bearing mineral phases, such as feldspars, which would appear to contain little trace metals. These findings further support the hypothesis that a significant proportion of metals would be associated with the silicate mineral phases in both types of sediments, and could thereby be considered inert or non-reactive relatively to early diagenetic processes. This metal fraction does contribute to the total content but is of no ecotoxicological interest. Therefore, trace metal normalization with total recoverable aluminum could be an interesting alternative to fixed, unique-value criteria in the process of sediment quality assessment, as shown in Figure 7.

Background levels and implications for sediment quality assessment

Because of their elevated Cr and Ni concentrations, postglacial marine clays have revealed themselves to be problematic in terms of implementing sediment quality assessment criteria. The toxicity of this particular sediment is currently being investigated by means of various bioassays to complement the results of the present study and to better define the nature and potential ecotoxicological impact of the sediment. Scientifically based tools for distinguishing postglacial marine clays within dredged materials could, therefore, become necessary if particular management decisions should so warrant.

Statistical comparisons between postglacial marine clays and St. Lawrence River sediments were performed on the data gathered in the present study. The variability observed between the 2 types of sediment was checked against the variability found in sediment core samples collected from different locations. Results from this test were then applied to all the sediment sampled in the present study in order to assure their validity at a more the global scale. Seventeen out of the 32 parameters analyzed showed average levels significantly different within both types of sediments (Table 6). This result suggested that the consideration of the analyzed variables on an individual basis could efficiently allow sediment classification. The discriminating analysis performed on these 17 variables in sediment samples retrieved from the same core identified Al as a successful separation instrument. Accordingly, sediment samples presenting total recoverable Al concentrations above 26,900 μg/g would be classified as postglacial marine clays, whereas total recoverable Al concentrations in St. Lawrence river preindustrial sediments should be lower than this value. When confronted to all the other sediment samples analyzed for the present study, this simple rule allowed to positively identify 98% of the postglacial marine clays samples.

CONCLUSION

Chromium, Cu, and Ni concentrations in analyzed samples of postglacial marine clays strongly exceeded, by a majority proportion, the MET and TET thresholds levels of the interim sediment quality criteria. The present study, however, showed that where metal concentrations in postglacial marine clays exceed guidelines, the potentially bioavailable fractions can be low and contrarily indicates low ecological risks. Analytical results demonstrate that these metals are unlikely to be bioavailable to the benthic fauna. Instead, these metals appear to be mostly incorporated into ferromagnesian aluminosilicates and could thereby be considered inert or nonreactive relatively to early diagenetic processes.

Because the 2nd and 3rd threshold levels of the interim criteria were established from data acquired essentially with the use of a total recoverable leaching procedure, it would be judicious and more reliable that the NET, which corresponds to the natural level of contaminants in the environment, be developed with similar analytical methods. As criteria development now tends to involve not only chemical analyses but also ecotoxicological assessments, the use of leaching reagents allowing the determination of potentially bioavailable contaminants in sediments, rather than their total content, becomes more promising. Although the total recoverable extraction probably still represents a relatively aggressive chemical extraction, it nevertheless solubilize metals associated with phases more reactive than are the most inert silicates such as feldspar and quartz. The metal fraction so extracted may be not entirely of ecotoxicological relevance but would remain an acceptable, protective approach for sediment quality assessments. Special attention should, however, be given for Zn and Li analysis with this extraction method in order to avoid overestimations of their concentrations.

Figure Figure 6..

Relation between chromium, copper, and nickel with total (a) and total recoverable (b) lithium.

Because of their specific geochemical characteristics, postglacial marine clays should be better considered differently than are sediments in regard of the sediment quality criteria implementation. If special management criteria were to be developed for the postglacial marine clays, in light of these results, the use of Al concentrations in sediments, in conjunction with other physical characteristics, would reveal to be an interesting tool to identify this specific material. However, normalization approaches would also allow better discrimination of natural sediments from contaminated sediments and help to define sediment management disposition for postglacial marine clays dredged in the St. Lawrence river seaway. Normalization of metal concentrations with total recoverable Al concentrations would be an interesting alternative to specific criteria.

Figure Figure 7..

Determination of natural and anthropogenic sediments using trace metal normalization with total recoverable aluminum.

Because toxicological bioassays on these clay materials are warranty for a better environmental management, a complementary laboratory study is currently being undertaken to confirm the predicted low metal bioavailability.

Table Table 6.. Probabilities from analysis of variance on all the sediment sampled in the present study and analyzed with total recoverable leaching procedure
ParametersProbabilities
  1. a Lack of data.

Sodiuma<0.0001
Aluminum, copper, zinc, nickel, magneisum, rubidium, vanadium<0.001
Lithium, pottasium, gallium, cobalt, manganese, lead, barium, chromium, titanium, uranium, lanthanum, humidity<0.01
Iron, beryllium, strontium, silt + clay, sand<0.05
Clay, silt, calcium, mercury, arsenic, cadmium, organic carbon>0.05

Acknowledgements

This project was funded by the 3rd phase of the St. Lawrence Action Plan, a coagreement between Quebec federal and provincial governments initiated in 1988. The project was associated to the sustainable navigation component, which includes aspects such as sustainable management of dredging operations in the St. Lawrence River, as well as the revision of the interim criteria for quality assessment of St. Lawrence river sediments. The authors wish to thank M Arseneau, S Lepage, and L-F Richard for their help in sampling and P Collin and P Gagnon for their assistance in the data treatment.

Ancillary