Prepared by: Michael A. Zarull, John H. Hartig, and Lisa Maynard
Sediment Priority Action Committee
Great Lakes Water Quality Board
IV. CONTAMINATED SEDIMENT AND THE AQUATIC ENVIRONMENT
In the Great Lakes, as in many aquatic systems, a considerable mass of persistent contaminants can be found in the bottom sediment. The accumulation of contaminants in the sediment at levels that are not rapidly lethal may result in long-term, subtle effects to the biota by direct uptake or through the foodweb. The cycling and bioavailability of sediment-associated contaminants in aquatic systems over both short and long time frames are controlled by physical, chemical, biological, and geological processes.
Physical processes affecting sediment contaminant distribution include mechanical disturbance at the sediment-water interface as a result of bioturbation, advection and diffusion, particle settling, resuspension, and burial. Some examples of significant geological processes affecting contaminant distribution and availability include weathering or mineral degradation, mineralization, leaching, and sedimentation. Chemical processes such as dissolution and precipitation, desorption, and oxidation and reduction can have profound effects, as well as biological processes such as decomposition, biochemical transformation, gas production and consumption, cell wall and membrane exchange/permeability, food web transfer, digestion, methylation, and pellet generation. In addition, the fundamental differences in physical, chemical, and biological properties and behavior of organic versus inorganic substances (metals, persistent organics, organo-metals, and nutrients) suggests the need for a more detailed knowledge of the area and the relative importance of these processes prior to completing an assessment of impact or planning remedial measures. Details of the major processes and their effects on contaminant cycling and movement can be found in Forstner and Whittman (1979), Salamons and Forstner (1984), Allan (1986), and Krezovich et al. (1987); however, it is important to explore the factors that affect bioavailability and uptake of contaminants, as well as the likely, quantifiable consequences of bioaccumulation.
The rate and mechanism of contaminant uptake from sediment by bottom-dwelling organisms can vary considerably among species, and even within species. Factors such as feeding ecology of the organisms, their developmental stage, season, behavior, and history of exposure affect contaminant uptake and body burdens. As well, different routes of uptake (soluble transfers versus contaminated food) can also be expected to affect tissue levels.
Experiments with organochlorine pesticides have yielded conflicting results on the relative significance of diet versus aqueous uptake. Within individual studies, available data on sediment- based bioconcentration factors for various organisms show a wide variation among species for a specific contaminant (Roesijadi et al. 1978a; 1978b). Accumulation of both organic and metal contaminants can be passive due to adsorption onto the organism, or it can be an active process driven through respiration. "Case-dwelling" species of benthic invertebrates have been thought less susceptible to contaminants than "free-living" organisms since the bioconcentration factors (BCF) have been found to be quite different for metals like copper and zinc. Similar differences have been found for oligochaete and amphipod tissue concentrations for PCBs and hexachlorobenzene.
Sediment type can profoundly influence the bioavailability of sediment-sorbed chemicals. Many researchers have reported an inverse relationship between chemical availability and sediment organic carbon content (Augenfield and Anderson 1982; Adams et al. 1983). There also appears to be a smaller, not as well defined relationship between sediment particle size and chemical availability. In fine-grained sediment, this is most likely due to the increased surface area available for adsorption and the reduced volume of interstitial water. Chemicals sorbed to suspensions of organic particles (both living, such as plankton, and non-living) may constitute sources of exposure for filter-feeding organisms and may be important in deposition. This pathway may be significant, as these organisms have been shown to accelerate the sedimentation processes by efficiently removing and depositing particles contained in the water column.
Several water quality conditions influence bioaccumulation of contaminants: temperature, pH, redox, water hardness, and physical disturbance. In addition, metals in mixtures may also compete for binding sites on organic molecules, resulting in antagonistic effects (e.g., cadmium and zinc, silver and copper).
The biological community itself can strongly influence the physical-chemical environment in the sediment, and in turn, affect the bioavailability of contaminants. For example: primary productivity influences the pH, which can influence metal chemistry; sulphate reduction by bacteria facilitates sulphide formation; the reduction of oxygen by organisms and their activities to anoxia affects redox conditions, and with it, metal redox conversion; the production of organic matter that may complex with contaminants; bioturbation influences sediment-water exchange processes and redox conditions; and methylation of some metals such as mercury.
Water based, BCFs indicate that benthic invertebrates generally accumulate to higher concentrations than do fish. This may be attributed to the greater degree of exposure of the benthic invertebrates at the sediment-water interface than fish. Biomagnification occurs when contaminant concentrations increase with successive steps in the trophic structure. However, well defined trophic levels may not exist in the aquatic ecosystem under examination, especially ones experiencing (or that have experienced) anthropogenically generated loadings of various contaminants. In addition, individual species may occupy more than one trophic level during the life cycle. These factors not only complicate process and exposure understanding, they also complicate monitoring program designs necessary to document improvement after remediation has taken place.
Metals, in their inorganic forms, do not appear to biomagnify appreciably in aquatic ecosystems; however, methylated forms of metals, like mercury, do biomagnify. Most persistent toxic organics demonstrate biomagnification to lesser or greater degrees; however, it appears that biomagnification is not as dramatic within aquatic food chains as terrestrial ones. Also, it appears that where the phenomenon does occur, the biomagnification factors between the lowest and highest trophic levels are usually less than one order of magnitude (U.S. Army Corps - Waterways Experiment Station 1984).
Ecological Effects of Contaminated Sediment
It was commonly assumed that chemicals sequestered within sediment were unavailable to biota, and therefore posed little threat to aquatic ecosystems. Although the laboratory and field studies are not overwhelming in number, both the risk and the impairment to organisms, including humans, have been conclusively established. Biota exposed to contaminated sediment may exhibit increased mortality, reduced growth and fecundity, or morphological anomalies. Studies have also shown that contaminated sediment can be responsible for mutagenic and other genotoxic impairments (Lower et al. 1985; West et al. 1986). These effects are not restricted to benthic organisms - plankton, fish, and humans are also affected both from direct contact and through the food chain.
Nuisance algal growth and nutrient relationships in lakes are well documented, with phosphorus being cited as the limiting nutrient. Some phosphorus is released during spring and fall lake circulation in dimictic lakes. In shallow, polymictic lakes, sedimentary phosphorus release may be more frequent, creating greater nuisance problems with the infusion of nutrients to overlying water, especially during summer recreational periods. This influx of nutrients usually results in abundant, undesirable phytoplankton growth, reducing water transparency, increasing color, and in severe cases, seriously depleting dissolved oxygen and potentially leading to fish kills. In addition, phytoplankton may be adversely impacted by contaminant-laden particulate matter.
Nau-Ritter and Wurster (1983) demonstrated that PCBs desorbed from chlorite and illite particles inhibited photosynthesis and reduced the chlorophyll - a content of natural phytoplankton assemblages. In a similar study, Powers et al. (1982) found that PCBs desorbed from particles caused reduced algal growth as well as reduced chlorophyll production. The time course for desorption and bioaccumulation appears to be quite rapid, with effects being documented within hours after exposure (Harding and Phillips 1978). The rapid transfer of PCBs and other xenobiotic chemicals from particulate material to phytoplankton has significant ramifications because it provides a mechanism for contaminants to be readily introduced to the base of the food web.
The detrimental effects of contaminated sediment on benthic and pelagic invertebrate organisms have been demonstrated in several laboratory studies. Prater and Anderson (1977a; 1977b), Hoke and Prater (1980), and Malueg et al. (1983) have shown that sediment taken from a variety of lentic and lotic ecosystems was lethal to invertebrates during short-term bioassays. Tagatz et al. (1985) exposed macrobenthic communities to sediment-bound and water-borne chlorinated organics, and found similar reductions in diversity to both exposures. Chapman and Fink (1984) measured the lethal and sublethal effects of contaminated whole sediment and sediment elutriates on the life cycle of a marine polychaete, and found that both sources were capable of producing abnormalities, mortalities, and reduced fecundities in larval and adult worms. The biotransformation of sediment-derived benzo[a]pyrene has been shown to result in the formation of potentially mutagenic and carcinogenic metabolites in depositional feeding amphipods (Reichert et al.1985). Other sublethal effects may be more subtle; for example, infaunal polychaetes, bivalves, and amphipods have been shown to exhibit impaired burrowing behavior when placed in pesticide-contaminated sediment (Gannon and Beeton 1971; Mohlenberg and Kiorboe 1983). Some observations have linked contaminants in sediment with alterations in genetic structure or aberrations in genetic expression. Warwick (1980) observed deformities in chironomid larvae mouthparts, which he attributed to contaminants. Wiederholm (1984) showed similar deformities in chironomid mouthparts ranging from occurrence rates of less than 1% at unpolluted sites (background) to 5-25% at highly polluted sites in Sweden. Milbrink (1983) has shown setal deformities in oligochaetes exposed to high sediment mercury levels.
Fish populations may also be impacted by chemicals derived from contaminated sediment. Laboratory studies have shown that fathead minnows held in the presence of contaminated natural sediment may suffer significant mortalities (Prater and Anderson 1977a, 1977b; Hoke and Prater 1980). Morphological anomalies have also been traced to contaminated sediment associations with fish. Malins et al. (1984) found consistent correlations between the occurrence of hepatic neoplasms in bottom-dwelling fish and concentrations of polynuclear aromatic hydrocarbons in sediment from Puget Sound, Washington. In addition, Harder et al. (1983) have demonstrated that sediment-degraded toxaphene was more toxic to the white mullet than to the non-degraded form. These studies illustrate the potential importance of sediment to the health and survival of pelagic and demersal fish species, but do not necessarily indicate a cause and effect relationship. While we can expect that fish will be exposed to chemicals that desorb from sediment and suspended particles, the relative contributions of these pathways to any observable biological effects are not obvious. Instead, laboratory bioassays and bioconcentration studies are often required as conclusive supporting evidence. The Elizabeth River, a subestuary of the Chesapeake Bay, is heavily contaminated with a variety of pollutants, particularly PAHs. The frequency and intensity of neoplasms, cataracts, enzyme induction, fin rot, and other lesions observed in fish populations have been correlated with the extent of sediment contamination. In addition, bioaccumulation of these same compounds in fish and resident crabs was also observed. However, essential laboratory studies were not conducted to establish contaminants in sediment as the cause of the observed impairments (U.S. Environmental Protection Agency 1998b).
There have been few examples of direct impacts of contaminated sediment on wildlife or humans. Bishop et al. (1995; 1999) found good correlations between a variety of chlorinated hydrocarbons in the sediment and concentrations in bird eggs. They felt this relationship indicated that the female contaminant body burden was obtained locally, just prior to egg laying. Other studies by Bishop et al. indicated a link between exposure of snapping turtle (Chelydra s. serpentina) eggs to contaminants (including sediment exposure) and developmental success (Bishop et al. 1991; 1998). Other investigations of environmentally occurring persistent organics have shown bioaccumulation and a range of effects in the mudpuppy (Necturus maculosus) (Bonin et al. 1995; Gendron et al. 1997). In the case of humans (Homo sapiens) there is only anecdotal evidence from cases like Monguagon Creek, a small tributary to the Detroit River, where incidental human contact with the sediment resulted in a skin rash. For the most part, assessments of sediment-associated contaminant impacts on the health of vertebrates (beyond fish) are inferential. This approach is known as risk assessment, and it involves hazard identification, toxicity assessment, exposure assessment, and risk characterization (National Academy of Sciences 1983).
Superfund risk assessments, which are aimed at evaluating and protecting human health, are designed to evaluate current and potential risks to the "reasonably maximally exposed individual" (U.S. Environmental Protection Agency 1989). Both cancer and non-cancer health effects for adults and children are evaluated. Data for the evaluation include concentrations of specific chemicals in the sediment, water column, and other media that are relevant to the potential exposure route. These routes of exposure may include: ingestion of contaminated water, inhalation of chemicals that volatilize, dermal contact, and fish consumption. The media-specific chemicals of potential concern are characterized based on their potential to cause either cancer or non-cancer health effects, or both. Once the "hazards" have been identified, the prescribed approach is continued to include toxicity evaluation, exposure assessment, and risk characterization. All of this leads to a potential remedial action, which itself follows a set of prescribed rules.
"Ecological risk assessment (ERA) is the estimation of the likelihood of undesired effects of human actions or natural events and the accompanying risks to nonhuman organisms, populations, and ecosystems" (Sutter 1997). The structure of ERA is based on human health risk assessment (HHRA), but it has been modified to accommodate differences between ecological systems and humans. "The principal one is that, unlike HHRA, which begins by identifying the hazard (e.g., the chemical is a carcinogen), ERA begins by dealing with the diversity of entities and responses that may be affected, of interactions and secondary effects that may occur, of scales at which effects may be considered, and of modes of exposure" (Sutter 1997). Risk characterization is by weight of evidence. Data from chemical analyses, toxicity tests, biological surveys, and biomarkers are employed to estimate the likelihood that significant effects are occurring, or will occur. The assessment requires that the nature, magnitude, and extent of effects on the designated assessment endpoints be depicted.
It is apparent that rarely is the relationship between a particular contaminant in the sediment and some observed ecological effect straightforward. Physical, chemical, and biological factors are interactive, antagonistic, and highly dynamic. These things often preclude a precise quantification of the degree of ecological impairment or effect attributable to a contaminant present in the sediment, and therefore, the degree of ecological improvement or benefit that can be achieved through remediation. Precision in quantifying impairment, remediation, and recovery is always improved through a better understanding of both the specifics of ecosystem functioning, as well as the behavior of the chemical(s) of concern in that particular ecosystem. Although a basic understanding of aquatic ecosystem function and chemical fate is generally available, it is also evident that systems appear to be sufficiently unique and our understanding sufficiently lacking. Therefore, an adaptive management approach is the prudent course to follow. This requires a much tighter coupling of research, monitoring, and management in every case to develop quantifiable, realistic goals and measures of success to achieve them.
Sediment Remediation and Ecological Improvements
Sediment removal has been used as a management technique in lakes as a means of deepening a lake to improve its recreational potential, to remove toxic substances from the system, to reduce nuisance aquatic macrophyte growth, and to prevent or reduce the internal nutrient cycling which may represent a significant fraction of the total nutrient loading (Larsen et al. 1975). Below are some examples of the removal of sediment contaminated by a nutrient (phosphorus), a metal (mercury), and a persistent toxic organic compound (PCBs) from lakes, rivers, and embayments outside the Great Lakes Basin.
Lake Trummen, Sweden, is one of the most thoroughly documented dredging projects in the world. An evaluation of the effectiveness of the dredging, whose main purpose was to reduce internal nutrient cycling and enrichment through sediment removal, took place over a twenty year plus time frame.
Lake Trummen, with a surface area of approximately 1 km2, a drainage basin of some 12 km2, and a mean depth of 2 m, was originally oligotrophic; however, it became hypertrophic after receiving both municipal and industrial discharges over a long period of time. In order to rectify the problems, both municipal and industrial waste effluents were curtailed in the late 1950s; however, the lake did not recover. In the late 1960s, extensive research was undertaken, resulting in the removal of some 400,000 m3 of surface sediment (the top meter, in two 50 cm dredgings) from the main basin in 1970 and 1971.
Bengtsson et al. (1975) indicated that post-dredging water column concentrations of phosphorus and nitrogen decreased drastically and that the role of the sediment in recycling nutrients was minimized. Phytoplankton diversity increased substantially, while at the same time their productivity was significantly reduced. The size distribution of phytoplankton also shifted to much smaller cells, and water column transparency more than tripled. The troublesome blue- green algal biomass was drastically reduced, with some nuisance species disappearing altogether (Cronberg et al. 1975). Conditions in the lake had improved to such a degree by the mid 1970s that an additional research and management program was undertaken on the fish community. From the late 1960s throughout the 1980s, an extensive monitoring program was maintained. By the mid 1980s, this program not only documented a deterioration in water quality, but also the ecological response to the change; and it also helped to ascertain that the changes were due to increased nutrient inputs from the atmosphere and the surrounding drainage basin.
Similar sediment removal projects have been conducted in other areas: Vajgar pond in the Czech Republic, Lake Herman in South Dakota, and Lake Trehorningen in Sweden, just to name a few. The latter named project is of particular note, because although there were significant decreases in the water column concentrations of phosphorus, it remained too high to be algal growth limiting. As a result, algal biomass remained the same as before the dredging was undertaken. This illustrates the importance of having a good understanding and quantification of ecological processes prior to undertaking a remediation project. In addition, Peterson (1982) notes that through the early 1980s there was little evidence to support the effectiveness of sediment removal as a mechanism of ecological remediation. This lack of supporting research and monitoring data continues to be an obstacle to establishing the effectiveness of sediment cleanups.
Minamata Bay, located in southwestern Japan, is the site of one of the more notorious cases of metal pollution in the environment, and its subsequent impacts on human health. A chemical factory released mercury contaminated effluent into the Bay from 1932 to 1968. In addition to contaminating the water and sediment, methylated mercury accumulated in fish and shellfish. This resulted in toxic central nervous system disease among the individuals who ate these fisheries products over long periods of time. In 1973, the Provisional Standard for Removal of Mercury Contaminated Bottom Sediment was established by the Japanese Environmental Agency. Under this criterion, it was estimated that some 1,500,000 m3 of sediment would need to be removed from an area of 2,000,000 m2. Dredging and disposal commenced in 1977 along with an environmental monitoring program to ensure that the activities were not further contaminating the environment. Monitoring included measuring turbidity and other water quality variables, as well as tissue analysis of natural and caged fish for mercury residues. Dredging was completed in 1987, and by 1988 the sampling surveys provided satisfactory evidence that the goals had been achieved. Results of the ongoing monitoring showed that no further deterioration of water quality or increase in fish tissue concentration was occurring. By March of 1990, the confined disposal facility received its final clean cover. The total cost for the project was approximately $40-$42 million U.S. dollars.
Post-project monitoring provided clear evidence of a reduction in surficial sediment concentrations of mercury to a maximum of 8.75 mg/kg and an average concentration of below 5 mg/kg (national criterion is 25 mg/kg) (Ishikawa and Ikegaki 1980; Nakayama et al. 1992; Urabe 1993; Hosokawa 1993; Kudo et al. 1998). Mercury levels in fish in the bay rose to their maximum between 1978 and 1981, after the primary source had been cut off and some dredging had begun. Tissue concentrations declined slightly as dredging continued; however, they did fluctuate considerably. Fish tissue levels did finally decline below the target levels of 0.4 mg/kg in 1994, some four years after all dredging activity had ceased (Nakayama et al. 1996). These results demonstrate that mercury in the sediment continued to contaminate the fish and that removal or elimination of that exposure was essential for ecological recovery to occur. It also demonstrates that some impact (increased availability and increased fish tissue concentrations) could be associated with the dredging activity, and that a significant lag time from the cessation of remediation activity was necessary for the target body burdens to be achieved.
Persistent Toxic Organic Substances
During a 30 year period ending in 1977, at least 1.1 million pounds of PCBs were discharged into the Hudson River, New York, from two General Electric capacitor manufacturing plants located in Fort Edward and Hudson Falls. PCBs contaminated the water, sediment, and biota throughout a 320 km section of the Hudson River. Large-scale surveying and monitoring programs were begun in the mid 1970s to determine the extent of contamination, and to assist in the development and planning of remedial options. Activities including the reduction of PCB levels in the discharge, the dredging for navigational purposes of some 153,000 m3 of contaminated sediment, and the removal and stabilization of contaminated river bank sediment were conducted between 1977 and 1978.
In 1976, because of the concern over the bioaccumulation of PCBs in fish and other aquatic organisms and their subsequent consumption by people, the State of New York banned fishing in the Upper Hudson River and also banned commercial fishing of striped bass and several other species in the Lower Hudson River. The control of the discharge produced declines in the PCB levels in water, sediment, and fish tissue between 1977 and 1981. Subsequently, PCB levels in fish, which remain the impetus from remediation, have declined at a slower rate, but still persist at levels that cause the continuation of the fish consumption prohibitions and advisories.
U.S. EPA made an interim "no action" decision for the PCB contaminated sediment in 1984. The agency has been conducting a reassessment of its 1984 decision since 1990. In August 1995, the Upper Hudson River was re-opened to fishing, but only on a catch and release basis.
These few examples show that considerable ecological benefits can be obtained from the remediation of contaminated sediment. Surprisingly, the best documented ecological changes are associated with actions relating to nutrient problems, generally in small lakes and ponds and in areas of low human population density, and usually the least costly remediations. Since affiliated research and monitoring has been so lacking, it has been difficult to evaluate the overall success of sediment remediation in a general sense, i.e., to reasonably transfer lessons learned and recommendations on what things are still essential to know, and to achieve cost-effective and essential ecological remediation.
In some cases, even those projects with a great deal of pre-remediation research and monitoring, both unforseen results, as well as disappointing results, were obtained. This reinforces the need to see the approach to sediment remediation as an "adaptive management" phenomenon.