Assessment and Remediation of Contaminated Sediments (ARCS) Program
Table of Contents
- Chapter 1
- Chapter 2
- Chapter 3
- Chapter 4
- Chapter 5
- Chapter 6
- Chapter 7
- Chapter 8
- Chapter 9
- Chapter 10
- List of Tables
- List of Figures
Assessment Guidance Document
US Environmental Protection Agency. 1994. ARCS Assessment Guidance Document. EPA 905-B94-002. Chicago, Ill.: Great Lakes National Program Office.
Table Of ContentsASSESSMENT OF BENTHIC INVERTEBRATE COMMUNITY STRUCTURE
- EXPERIMENTAL DESIGN
- METHODS FOR SAMPLE COLLECTION
- DATA ANALYSIS
- THE ARCS APPROACH
- SUMMARY AND RECOMMENDATIONS FOR FUTURE STUDIES
Contaminated sediments are a major source of pollution in the United States and represent a potential threat to all components of aquatic ecosystems (Sorensen et al. 1977; Landrum and Robbins 1990). Sediments are a repository for organic and inorganic contaminants that can accumulate to high concentrations (Shimp et al. 1971; Oschwald 1972; Medine and McCutcheon 1989). Benthic invertebrates are closely associated with surficial sediments and therefore are continuously exposed to contaminants in the sediments.
Since aquatic ecosystems are composed of interdependent trophic levels, it generally is not appropriate to study individual components of an ecosystem when making assessments of sediment toxicity (Burton 1991). Complete ecological assessments of sediment toxicity usually require the use of resident biota as indicators of sediment quality. For the assessment to be successful, closely integrated biological, chemical, and physical data are required. Since sediments tend to integrate historical water quality conditions, the spatial and temporal distribution of resident organisms can reflect the degree to which chemicals in the sediments are toxic. Field surveys of benthic invertebrates provide an essential component of biological assessments of the toxicity associated with contaminated sediments. These surveys have several advantages: 1) indigenous benthic organisms complete all or most of their life cycles in the aquatic environment and serve as continuous monitors of sediment quality, 2) many benthic invertebrates living in sediments are relatively sedentary and are therefore representative of local conditions, 3) macroinvertebrates are relatively easy to collect and are generally abundant across a broad array of sediment types, 4) a field assessment of natural populations can provide a screening-level evaluation of potential sediment contamination, and 5) results of an assessment of indigenous populations are usually biologically interpretable, which allows resource injuries to be quantified in a manner more easily understood by managers, regulators, and the general public (Cook 1976; Pratt and Coler 1976; Davis and Lathrop 1989).
This chapter reviews the methods used to evaluate the response of benthic invertebrate communities to contaminated sediments and makes recommendations based on information gained during quantitative benthic invertebrate surveys conducted simultaneously with sediment evaluations under the ARCS Program (Canfield et al. 1993). The objective of the surveys was to describe the species distributions and relative abundances of benthic invertebrates and to interpret these in light of the chemical and physical characteristics of the sediments. This information, when analyzed in conjunction with the results of sediment toxicity tests (see Chapter 6), will provide a more complete representation of the effects of in situ contaminants on benthic invertebrate communities. As many as 10 stations were sampled from each of three priority AOCs: the Buffalo River in New York (Figure 1-1), Indiana Harbor in Indiana (Figure 1-2), and the Saginaw River in Michigan (Figure 1-3). Benthic community evaluations described in Canfield et al. (1993) include descriptions of estimated abundances for the total benthic community and individuals identified to the lowest possible taxon at each AOC, comparisons between the characteristics of benthic communities and the concentrations of sediment contaminants, evaluation of the prevalence of deformities in chironomids (midges), and analyses of the sources of variability in collecting representative benthic samples with a Ponar grab sampler.
Selected ARCS Program data sets from Canfield et al. (1993) are used in this chapter to evaluate 1) the usefulness of different indices for characterizing benthic invertebrate communities in contaminated sediments, 2) the numbers of samples needed to achieve a confidence level of 95 percent for estimated sample means, 3) the usefulness of statistical tests for evaluating the effects of sediment contamination on benthic invertebrate communities, and 4) key considerations for conducting future studies of benthic invertebrate communities in contaminated sediments. Data from Swift et al. (in prep.) are also used to evaluate the usefulness of artificial substrates for evaluating the effects of contaminants on benthic invertebrate communities.
The first step in conducting an evaluation of benthic invertebrate communities is the development of an appropriate experimental design. An inappropriate experimental design can be a major source of error in the resulting data (Thornton et al. 1982; La Point and Fairchild 1992). There are many factors that need to be considered when sampling contaminated sediments for benthic invertebrates that differ from the considerations required for sampling sediments for toxicity testing. Benthic invertebrate distributions are strongly influenced by abiotic factors in the absence of contaminants (Resh 1979; Pettigrove 1990; La Point et al. 1984; La Point and Fairchild 1992) and, in some cases, the effects of contaminants can be masked by effects due to abiotic factors. Important abiotic characteristics (e.g., sediment grain size, sediment organic content, sediment nutrient content, water quality, current velocity, water depth) at a study site should therefore be evaluated so that the potential confounding effects of these characteristics can be accounted for when the data are analyzed and interpreted. This holds true whether the intent of the project is to make comparisons between upstream and downstream areas, between different aquatic systems (i.e., different rivers or lakes), or between seasons.
When assessing benthic invertebrate communities for changes in community structure, it is critical to select appropriate reference sites with which the benthic invertebrate communities at study sites can be compared (Davis and Lathrop 1989; La Point and Fairchild 1992). Ideally, a reference site should be unaffected or minimally affected by anthropogenic influences. Since a completely unaffected system is difficult or impossible to find, it is usually considered acceptable to use sites that are considerably less contaminated than the study site (Chapman 1986). However, because low contaminant concentrations in water and sediment can sometimes affect benthic invertebrate communities, caution should be used when comparing results with a reference site that has contaminant concentrations higher than in pristine areas. A reference site should also have physical and chemical characteristics of both water and sediment that are similar to the study site to account for the potential effects of those characteristics on benthic communities at the study site.
Many studies have evaluated the number of replicate samples required to provide adequate assessments of benthic invertebrate communities and to allow cause-and-effect predictions to be made from the data (Elliott 1977; Green 1979; Resh 1979; Barton 1989). Many of these studies suggest that a sufficient number of replicate samples should be taken so that the among-sample coefficient of variation for all invertebrates is less than 50 percent (Davis and Lathrop 1989). To determine the number of benthic invertebrate samples that should be collected, it is recommended that a preliminary survey be conducted of the study areas. This is done to qualitatively identify the taxa that will be encountered and the relative abundances of those taxa at each station. Depending on the types of taxa collected, the methods used to collect benthic invertebrates may need to be modified to more effectively sample the benthos.
Although the optimum number of samples may be determined when designing a particular study, 3-5 replicate samples per station are usually collected in most studies. The reason for collecting 3-5 replicate samples is based primarily on funding and personnel constraints, which limit the processing of large numbers of samples. Although the collection of a smaller number of replicate samples may not invalidate the benthic invertebrate data for making community assessments, investigators should interpret such data with caution if the sample replicates are heterogeneous (i.e., abundance estimates will have high variance).
There are several books and articles pertaining to the proper design and conduct of benthic invertebrate surveys (Davis and Lathrop 1989; Plafkin et al. 1989; Hurlbert 1984; APHA 1985; Elliott 1977; Resh 1979; La Point and Fairchild 1992; Merritt and Cummins 1984; Klemm et al. 1990). These sources should be consulted during the planning stage of a benthic invertebrate survey to develop the optimum study design within the constraints of the financial limitations of the study.
The collection methods for benthic invertebrate samples depend on the type of habitat to be sampled (e.g., rocky substrate, fine-grained substrate, heavily vegetated areas) and the type of system (e.g., flowing or standing water) from which the samples are to be taken. Many of the benefits and limitations of different sampling devices have been described in previous publications (Resh 1979; Downing 1984; Klemm et al. 1990). In the ARCS Program, the sampling devices used were the 0.05-m2 Ponar grab sampler (Powers and Robertson 1967) and an artificial substrate sampler (Stauffer et al. 1976).
Grab samplers are designed to take discrete "bites" of the sediments that are representative of a fixed area and are therefore the preferred method for collecting sediments for the quantitative assessment of benthic infauna. The advantages and disadvantages of various grab samplers are discussed in Chapter 3, Sediment Sampling Surveys. Ponar grab samplers, both full-sized and petite, are among the best all-purpose samplers available for sampling unconsolidated sediments (Downing 1984). The Ponar grab sampler was chosen for the collection of sediment for the analysis of benthic invertebrates in the ARCS Program. It is recommended that a Ponar grab sampler be used for the collection of sediment in future studies of benthic invertebrate communities in the Great Lakes, because the benefits of this type of sampler outweigh any limitations.
Artificial substrate samplers have a long history of use in studies of benthic invertebrate communities in aquatic ecosystems. Artificial substrate samplers are designed to mimic natural substrates (e.g., gravel, cobble, small spaces) and to provide an easily quantified sampling unit. As with grab samplers, artificial substrate samplers can provide both qualitative and quantitative samples of benthic macroinvertebrates. Cairns (1982) and Klemm et al. (1990) review the advantages and limitations of artificial substrate samplers. General descriptions of the use of artificial substrate samplers in ecological and hazard assessments include those of Rosenberg and Resh (1982), Isom (1986), Ohio EPA (1987), and Klemm et al. (1990). Stauffer et al. (1976) and Swift (1985) describe the use of mesh-filled chicken baskets as artificial substrates.
In general, artificial substrate samplers primarily sample the epifaunal community, whereas grab samplers primarily sample the infaunal community. Artificial substrate samplers made of mesh are particularly good at collecting large numbers of animals because of the large number of interstitial spaces. Mesh artificial substrate samplers are a good alternative to grab samplers when collecting animals for tissue residue analyses.
Multiplate samplers (e.g., Hester-Dendy samplers) are designed to provide small spaces for benthic organisms to hide in. They are easy to make and readily available from commercial suppliers (e.g., Wildco). They have been used extensively in shallow water stream studies but less often in studies conducted in standing water. Like the mesh samplers, Hester-Dendy samplers tend to sample the epifaunal community.
Selection of the most appropriate samplers should be based on the objectives of the study as well as the depth of water at the site. The mesh samplers can be used easily in both shallow water and in deeper waters, whereas the Hester-Dendy samplers are primarily designed for use in shallower waters where wading is possible. Hester-Dendy samplers can be used in deeper waters, but they require the use of a heavy platform (i.e., steel plate or cement tile) as an attachment surface for the samplers to sit in the proper upright orientation on the bottom at the sampling site.
To determine which indices to evaluate in studying the effects of contaminants on the benthic invertebrate communities, it is necessary to determine the kind of information required for the planned data analysis. Some analytical methods are only amenable to a qualitative assessment of benthic invertebrate communities. However, the more desirable goal of a study is the quantitative assessment of the effects of the contaminants on benthic invertebrate communities. Several metrics are available for qualitative and quantitative assessments of benthic invertebrate communities (Merritt and Cummins 1984; Pennak 1989; Plafkin et al. 1989; Klemm et al. 1990). Community structure measurements can be divided into four broad categories: numbers of individuals and standing crop, multivariate analyses, diversity and similarity indices, and indicator organisms (IJC 1988). Benthic community health can be assessed by determining the structure (e.g., taxa richness), community balance (e.g., percent dominant taxa), and functional feeding group (e.g., percent scrapers, percent filter feeders) composition of the macroinvertebrate community (Plafkin et al. 1989) Most of these metrics are quantitative, although the use of indicator organisms tends to be more qualitative in nature. The most frequently used and simplest metrics are numerical abundance, percent composition of dominant taxa, and taxa richness (i.e., the number of taxa present). These indices have the advantage of being easily measured and are highly sensitive to contaminants and other anthropogenic perturbations (Sheehan and Winner 1984; IJC 1988; Van Hassel et al. 1988).
Multivariate analyses are frequently used for measuring patterns associated with benthic invertebrate distribution and relative abundance. Multivariate analyses typically fall into two categories: clustering and ordination methods (IJC 1988). In addition, other commonly used multivariate techniques include PCA, multiple regression analysis, and multiple ANOVA. There is considerable literature describing these multivariate techniques (Blackith and Reyment 1971; Poole 1974; Elliott 1977; Green 1979, 1980; IJC 1988). Multivariate approaches are used to answer questions relating to when and where a contaminant is affecting benthic communities. In highly contaminated areas, the large numbers of samples with zero abundances for many taxa often preclude the use of multivariate statistics.
Diversity and similarity indices have been widely used for assessing the impacts of contaminants (Peet 1974; Pielou 1977; Green 1979; Sheehan 1984; Klemm et al. 1990). In theory, communities that are unaffected by contaminants have higher diversity and communities affected by contaminants have lower diversity. The primary advantage of using diversity indices is that a large amount of data is reduced to a single number representing an entire community. However, the use of a single number can result in the loss of information that would be retained by the use of other statistical methods (IJC 1988). In addition, the reliance on a single index can produce misleading results. For example, use of the Shannon-Wiener diversity index for samples having few individuals and a relatively even distribution among several species could yield a high diversity value even if the study site is extremely contaminated (Green 1979). The use of diversity indices has decreased in recent years in favor of other indices such as numerical abundance, biomass, taxa richness, or composite or multimetric indices (Ohio EPA 1989; Plafkin et al. 1989). There is considerable literature describing diversity and similarity indices (Cairnes et al. 1982; Washington 1984; IJC 1988; Davis and Lathrop 1989; Plafkin et al. 1989; Klemm et al 1990).
Indicator species or communities are primarily qualitative indices. Some of the indices most widely used are the modified Hilsenhoff Biotic Index, the Ephemeroptera + Plecoptera + Trichoptera Index, the oligochaete/chironomid ratio, the percent contribution of dominant taxa, the community loss index, the Jaccard coefficient of community similarity, and the ratio of shredders to total abundance (IJC 1988; Plafkin et al. 1989; Klemm et al. 1990). The use of indicator species is based on a prior knowledge of the contaminant tolerances of various taxonomic groups. Some invertebrate taxa are known to be relatively tolerant to organic enrichment and chemical contamination (e.g., certain oligochaetes and chironomids), whereas other groups are characteristically intolerant (e.g., certain mayflies and stoneflies). Within these broad groups, it is important to be able to identify individuals to at least the generic level, because there are both tolerant and intolerant taxa within most of these groups (Resh and Unzicker 1975; Plafkin et al. 1989; Klemm et al. 1990). For example, within the family Chironomidae, the genus Chironomus is generally more tolerant of organic enrichment than the genus Polypedilum (Klemm et al. 1990). Within the genus Chironomus, there are differences in the tolerances of Chironomus riparius, which is considered extremely tolerant of organic enrichment, and Chironomus tuxis, which is considered intolerant of organic enrichment (Klemm et al. 1990).
The purpose of this section is to provide "lessons-learned" information using data from the ARCS Program. A general overview and the goals of the ARCS Program are described in detail in Chapter 1, Introduction, and Chapter 6, Evaluation of Sediment Toxicity. Additional details pertaining to the specifics of the benthic community survey are presented in this chapter. This section identifies the positive lessons learned from the benthic invertebrate surveys conducted in the ARCS Program, as well as those aspects of the studies that can be strengthened for future work.
Assessments of benthic invertebrate communities were conducted as part of the ARCS Program in three priority AOCs in the Great Lakes: the Buffalo River in New York (Figure 1-1), Indiana Harbor in Indiana (Figure 1-2), and the Saginaw River in Michigan (Figure 1-3). All three priority AOCs receive municipal and industrial wastewater discharges that contain a variety of organic and inorganic contaminants.
In the ARCS Program, benthic invertebrate samples were collected from each of the three priority AOCs. A total of 155 benthic grab samples (about 5 grabs per station and 10 stations per AOC) were collected. Artificial substrate samplers were deployed at five stations in the Buffalo River, four stations in Indiana Harbor, and six stations in the Saginaw River. To minimize potential disturbance of the sediments and associated invertebrates, all 155 benthic grab samples were collected before sediment samples were collected for chemistry and toxicity evaluations. The five replicate benthic grab samples at each station were collected within a 100-m2 area. Each benthic grab sample was sieved through a 500-um brass screen; site water was used for rinsing. Material retained by the sieve was rinsed into 500-mL glass jars and preserved with 10-percent buffered formalin.
Before sorting, samples were rinsed thoroughly with tap water to remove formalin and excess silt or mud. The rinsed samples were drained of excess water, returned to their original jars, and allowed to soak in 95-percent ethanol for at least 24 hours to facilitate extraction of any volatile chemicals. After the 24-hour soaking period, each sample was rinsed again with tap water to remove the ethanol and volatile chemicals. Each sample (except the samples from Indiana Harbor Station 10) was placed in a 4-L wide-mouth jar and agitated with tap water so that the invertebrates and lighter detrital material floated while the snails, clams, and heavier material remained on the bottom of the jar. Aliquots of the sample were removed from the jar to sort the benthic invertebrates from the debris. Aliquots were removed until the entire sample had been sorted. In the case of the samples from Indiana Harbor Station 10, only a portion (50 mL) of each sample was sorted and enumerated, because the abundance of oligochaetes was so high that enumeration of the entire sample could not be conducted in a timely manner. Sorting times ranged from approximately 3 to 20 hours per sample.
A binocular dissecting microscope with a magnifying power of 4x-12x was used to sort the samples. Organisms were sorted and enumerated into the following orders or families: Oligochaeta, Chironomidae, Bivalvia, Gastropoda, Ephemeroptera, Odonata, Plecoptera, Hemiptera, Megaloptera, Trichoptera, Coleoptera, Diptera (other than Chironomidae), Hirudinea, and Amphipoda. These samples were used to estimate macroinvertebrate numerical abundance (individuals/m2), species composition, and taxa richness. Taxonomic identifications were made by National Fisheries Contaminant Research Center (NFCRC) personnel using published taxonomic keys (Wiederholm 1983; Merritt and Cummins 1984; Pennak 1989; Thorp and Covich 1991). Oligochaetes and chironomids were mounted on slides for identification. Oligochaetes were identified to genus and species (when possible), and chironomids were identified to genus. Molluscs were identified to genus and species, and all other taxa were identified to the lowest practical taxon (usually genus and species).
Chironomid larvae were also examined for deformities in mouthpart structures. These deformities consisted of various types of asymmetry, missing teeth, extra teeth, fusion among various teeth, and labial separation, as described by several investigators (Saether 1970; Hamilton and Saether 1971; Hare and Carter 1976; Warwick et al. 1987; Warwick 1989). Individual chironomid larvae were mounted on slides and examined for deformities in the mentum (Orthocladinae and Chironominae) and ligula (Tanypodinae). The prevalence of mouthpart deformities was calculated as a proportion of the total number of chironomid larvae found at each station.
Artificial substrate samplers were constructed from 3M ® synthetic mesh and stainless-steel wire rotisserie chicken baskets (Stauffer et al. 1976). Each substrate consisted of five pieces of mesh (20 x 20 cm) folded in half and placed beside each other in a basket. The baskets were 26 cm in length, 17 cm in diameter, and 53 cm in circumference (Figure 7-1). The baskets were wired shut, and three baskets were wired to a cinder block at each sampling station. The baskets were connected to the cinder block with 2-m wires and were placed horizontally on the bottom near the cinder blocks. One end of the wire was attached to the cinder block and the other end was connected to a recognizable landmark on shore to facilitate retrieval of the artificial substrate samplers.
Artificial substrate samplers were deployed at five stations in the Buffalo River (October 1989), four stations in Indiana Harbor (August 1991), and six stations in the Saginaw River (June 1990), with deployment at each station lasting for 30 days. Each substrate sampler was placed on the surface of the sediment. Upon retrieval, the samplers were lifted carefully to the water surface and placed in plastic dish pans. The mesh substrate material was removed, placed in 4-L wide-mouth jars, and preserved in 4-percent buffered formalin. The water and sediments remaining in the dish pans were poured through a 250-um mesh sieve, and the retained material was preserved with the synthetic mesh.
The artificial substrate samples were rinsed in the laboratory through a 250-um mesh sieve, and each piece of mesh was then unraveled under water in a dish pan. Sediment and organisms were retained in the water, sieved (250-um mesh), and stored in 70-percent ethanol. The mesh was discarded. The entire preserved sample was placed in a pan for sorting at 4x-12x magnification. A subsample of 100 organisms was removed from each sample. Organisms were subsampled in approximate proportion to their relative abundance in the sample following the Rapid Bioassessment Protocols (Plafkin et al. 1989). At Indiana Harbor Station 3, each sample contained thousands of small (<500 um), recently settled zebra mussels (Dreissena sp.). A subsample of 100 invertebrates other than zebra mussels was picked to evaluate the other organisms present at that station. At all stations, organisms that were too rare to be included in the subsamples were recorded qualitatively.
Benthic invertebrate samples should be collected in a manner that provides the best possible estimate of benthic invertebrate community structure. To minimize potential disturbance of the sediments and associated invertebrates, all 155 benthic grab samples for the ARCS Program were collected before sediment samples were collected for chemistry and toxicity evaluations.
Samples were sorted in the laboratory by a number of technicians. To ensure that all samples were sorted with a similar efficiency, 1 of every 10 samples was randomly selected and resorted by a supervisor to confirm that the sample was sorted completely. If the number of invertebrates found during resorting was >=5 percent of the total number of invertebrates in the sample, all 10 samples in the lot were resorted in their entirety.
Two major elements of benthic invertebrate surveys can contribute to the variability associated with estimates of species distribution and abundance. The first is the variability associated with different field collection methods, and the second stems from inaccuracies in taxonomic identification. Although benthic community structure cannot be assessed for accuracy, precision was monitored. The precision associated with the collection of benthic invertebrate samples was evaluated by examining the five replicate grab samples collected at each station. The replicate samples were collected within a 100-m2 area at each station. The variance associated with field collection was evaluated using an ANOVA to identify the sources of variability.
The accuracy of taxonomic identifications was evaluated by having independent taxonomic experts outside of the NFCRC verify the identifications made by NFCRC personnel.
As described by Canfield et al. (1993), data were analyzed using appropriate parametric and nonparametric statistical tests (Snedecor and Cochran 1982; Statistical Analysis System 1988). Statistical analysis was performed using the Statistical Analysis System computer package for personal computers (Statistical Analysis System 1988). Relationships among the abundances of benthic invertebrates within and among AOCs were evaluated using a nested ANOVA (Snedecor and Cochran 1982). Comparisons between benthic invertebrate abundances and physical and chemical data were conducted using correlation and multivariate regression analyses. Unless otherwise specified, statements of statistical significance refer to significance at P<=0.05.
Raw data and summary statistics are presented in Canfield et al. (1993), Coyle et al. (1993), and Nelson et al. (1993). The data have also been entered into USEPA's Ocean Data Evaluation System database and have received quality assurance validation from the USEPA (Chapter 2). Selected data sets from the ARCS Program were used to evaluate the following questions:
- What is the benthic invertebrate community composition at each station?
- What is the benthic invertebrate community composition at each AOC?
- Is there a significant correlation between sediment contamination and benthic invertebrate community structure?
- Is there a relationship between larval chironomid mouthpart deformities, benthic invertebrate community composition, and physical and chemical sediment characteristics?
- What value do artificial substrate samplers have in the assessment of benthic invertebrate communities in relation to contaminated sediments?
- What are the sources of variability in the benthic invertebrate community analyses?
- What value do analyses of benthic invertebrate community structure have in the evaluation of contaminated sediment sites for potential remediation?
- How should evaluations of benthic invertebrate community structure be used in future Great Lakes studies?
Although a number of different metrics were used to evaluate the data collected for the ARCS Program, only the metrics that showed the best ability to discriminate among sites were used to evaluate the data. Those metrics were percent contribution of major taxa, comparisons between numerical abundances and species composition, comparisons between numerical abundances and sediment chemistry, prevalences of mouthpart deformities in larval chironomids, percent of total variance in abundance estimates, and evaluations of chironomid genera richness.
Benthic Invertebrate Numerical Abundance and Community Structure in Grab Samples
The estimates of the numerical abundances of benthic invertebrates in the ARCS Program were probably conservative (i.e., biased low) because of the sampling device used to collect the invertebrates (Resh 1979) and the size of the mesh used for sieving (Brinkhurst 1974; Resh 1979; Heushelle 1982). Ponar grab samplers are relatively heavy so that they can penetrate the sediment surface evenly and efficiently. Nevertheless, if the sediments are very soft, Ponar grab samplers can overpenetrate. However, a problem can occur as the sampler nears the sediment. A shock wave of displaced water can impact the sediment just before the sampler makes contact, causing small surface-dwelling animals to be pushed out of the way of the sampler (Flannagan 1970; Howmiller 1971; Howmiller and Beeton 1971; Milbrink and Wiederholm 1973). The effect of this type of disturbance is not easily quantified and was not addressed in these studies. However, any bias caused by this type of sampler should have been consistent across stations and AOCs.
All sediment grab samples were sieved through a 500-um mesh brass screen after field collection. Although this mesh size is adequate for separating benthic invertebrates from the sediments, many smaller organisms, such as the naidid worms and early life stages of midge larvae, pass through the sieve (Brinkhurst 1974; Mason et al. 1975; Resh 1979; Heushelle 1982). In retrospect, it is now considered more appropriate to sieve samples sequentially through 500-um and 250-um mesh screens to capture organisms that pass through a 500-um mesh screen (Burt et al. 1991; Brinkhurst 1992, pers. comm.). The 500-um mesh screen allows quick and efficient separation of larger organisms and debris from the smaller organisms and fine-grained sediments. If the 250-um mesh screen were used alone, the screen would tend to clog and sieving would be slowed considerably. Therefore, sequential sieving through 500-um and 250-um mesh screens is recommended for future studies of benthic invertebrates in Great Lakes sediments.
Comparisons Among AOCs: Benthic Invertebrate Numerical Abundance and Species Composition in Grab Samples
Benthic invertebrate samples from the Buffalo River exhibited a wide range of total numerical abundance across all stations (Table 7-1). Although oligochaetes were the most abundant organisms, several stations had a large number of chironomids, bivalves, and gastropods. Representatives from the orders Ephemeroptera, Odonata, Hemiptera, Trichoptera, Coleoptera, Diptera (other than Chironomidae), Hirudinea, and Amphipoda were rarely collected.
Indiana Harbor had a depauperate benthic invertebrate community. Except for two individual chironomids collected at Station IH-01-10, no other insects were present in the grab samples from Indiana Harbor (Table 7-2). Bivalve molluscs were rare, occurring only at three stations (IH-01-03, IH-01-04, IH-01-07). The bivalve genera (Musculium, Pisidium, and Sphaerium) found in Indiana Harbor are considered tolerant of organic enrichment (Carr and Hiltunen 1965; Fuller 1974; Bode 1988).
The benthic invertebrate samples from the Saginaw River exhibited a fairly narrow range of total numerical abundance (Tables 7-3 and 7-4). As with Indiana Harbor and the Buffalo River, oligochaetes were the most abundant organisms in the grab samples at all stations. Oligochaetes accounted for a higher percentage of the invertebrate communities across stations during the Saginaw River survey conducted in December 1989 than in the June 1990 survey. Chironomids were more abundant in the June 1990 survey than in the December 1989 survey. Although the oligochaetes and chironomids are present throughout the year in these sediments, it is likely that there are seasonal fluctuations in the abundance of both groups of organisms. These potential seasonal fluctuations should be considered when comparing the abundances of benthic invertebrate communities among different seasons.
Oligochaeta --Invertebrate communities dominated by the tubificid oligochaetes, to the exclusion of other invertebrate groups, are often indicative of organic enrichment (Brinkhurst et al. 1972; Brinkhurst and Cook 1974; Cook and Johnson 1974; Burt et al. 1991). In the Buffalo River, the oligochaetes were dominated by tubificids, including species (e.g., Limnodrilus hoffmeisteri) that are generally considered tolerant of organic enrichment and metal contamination (Table 7-5; Kennedy 1965; Brinkhurst et al. 1972; Burt et al. 1991). Limnodrilus hoffmeisteri was the most abundant oligochaete in the grab samples at all stations except Stations BR-01-01 and BR-01-10. The reasons for the lower abundances of this oligochaete species at those two stations are not clear, but may be a result of food or habitat preferences not being met there (Verdonschot 1989).
All of the tubificid genera present in Indiana Harbor are known to be tolerant to organic enrichment (Kennedy 1965; Brinkhurst et al. 1972). Limnodrilus hoffmeisteri, one of the most tolerant oligochaete species, was the most abundant species in the grab samples at all stations (Table 7-6).
The abundance of oligochaetes was lowest at Indiana Harbor Stations IH-01-07 (junction of Lake George Branch and Grand Calumet Branch) and IH-01-06 (main channel), and highest at Station IH-01-10. Sediments from Station IH-01-07 generally had the highest concentrations of metals and organic contaminants (Nelson et al. 1993), and Station IH-01-06 had the second highest concentrations of metals and organic contaminants. This indicates that the combined inputs of metals and organic contaminants from the Grand Calumet Branch and the Lake George Branch are affecting even the relatively contaminant-tolerant oligochaetes at these stations. Samples from these stations were also found to be among the most toxic sediments evaluated by the sediment toxicity tests (Coyle et al. 1993; Nelson et al. 1993). High concentrations of metals may reduce the abundance of oligochaetes by reducing the abundance of bacteria on which they feed.
The abundance of oligochaetes was extremely high at Station IH-01-10, approaching 1,000,000 individuals/m2 in individual grab samples. Although metals concentrations at that station were the second lowest of any of the stations sampled, the primary reason for the higher abundances of oligochaetes may be the high density of aquatic vegetation present at Station IH-01-10. Large amounts of vegetation were most likely present because Station IH-01-10 is on the upstream side of a low-clearance bridge, which tends to minimize the amount of disturbance caused by boat traffic. This vegetation probably retains decaying plant material, which enhances bacterial abundances.
In the Saginaw River, as in Indiana Harbor and the Buffalo River, the oligochaetes were dominated by the tolerant tubificids (Tables 7-7 and 7-8). Limnodrilus hoffmeisteri was more dominant in the samples from the December 1989 survey than in the samples from the June 1990 survey, in which the relative abundances of oligochaetes were distributed more evenly across several other species. These differences may be attributed to seasonal variability, spatial variability, or differing food sources.
Chironomidae--The chironomid community in the Buffalo River (Table 7-9) consisted primarily of genera known to be tolerant of organic enrichment (Hilsenhoff 1982, 1987; Beck 1977; Bode 1988; Klemm et al. 1990). The exception to this generalization was Tanytarsus at Station BR-01-10, which reportedly prefers less organically enriched environments (Krieger 1984). Chironomus, Procladius, Cryptochironomus, and Cricotopus are generally considered to be the most abundant chironomid genera in heavily contaminated environments (Cook and Johnson 1974; Krieger 1984). The first three of these genera were generally the most abundant chironomids collected from Buffalo River sediments.
Only two individual chironomids were collected in grab samples from Indiana Harbor. The larvae were identified as members of the genus Cricotopus, which is generally considered tolerant of organic enrichment and metals contamination.
The chironomid community in the Saginaw River was also comprised predominantly of tolerant genera (i.e., Chironomus, Cryptochironomus, Procladius) (Tables 7-10 and 7-11; Hilsenhoff 1982, 1987; Bode 1988). The exception was Tanytarsus at Station SR-03-08, which was present but in very low abundance. Procladius and Chironomus were the most frequently collected chironomid genera and are reported to be tolerant of organic enrichment and metals contamination (Cook and Johnson 1974; Krieger 1984; Klemm et al. 1990).
Mollusca--Several genera and species of Bivalvia (6 genera, 6 species) and gastropods (4 genera, 5 species) were present in the Buffalo River (Table 7-12). Two genera, Musculium and Pisidium, have been reported from organically enriched environments (Carr and Hiltunen 1965; Fuller 1974). Species of the genus Sphaerium may be somewhat tolerant to organic enrichment by virtue of being in the family Sphaeridae, which reportedly is tolerant of organic enrichment (Bode 1988; Plafkin et al. 1989).
Representatives from the family Unionidae were collected at only 4 of 10 stations in the Buffalo River and their abundances were always low. Although this pattern could indicate that these organisms are less tolerant than some of the Sphaeridae, members of the family Unionidae tend to be large and the Ponar grabs may not sample them efficiently.
Representatives from the genera Valvata and Bithynia have been reported to be somewhat tolerant of organic enrichment (Carr and Hiltunen 1965; Krieger 1984; Klemm et al. 1990). Even so, the occurrence of these genera was limited to 6 of 10 stations in the Buffalo River, and the abundances of these genera were usually low at most stations. The low abundances of those genera may be due to the toxic effects of metals and organic contaminants, the absence of sufficient grazing material, the lack of suitable habitats, or a combination of the above factors. Interestingly, two gastropod species, Valvata lewisi and Cincinnatia cincinnatiensis, which are reportedly uncommon in New York State (Jokinen 1992, pers. comm.), were the most abundant gastropod species in the grab samples from the Buffalo River stations. Other reported occurrences of these species in New York are in Oneida Lake, Seneca Lake, and several tributaries in the Oswego drainage basin. All of these areas are subject to relatively high levels of organic enrichment. This may indicate a tolerance for organic enrichment that allows these species to displace other gastropods and survive in such areas.
The only molluscs present in the grab samples in Indiana Harbor were fingernail clams (Sphaeridae). Musculium sp., Pisidium sp., and Sphaerium sp., were present at extremely low abundances (4/m2). Many of the grab samples from Indiana Harbor had large numbers of gastropod and bivalve shells. The affected stations were located in depositional areas and the shells may have been carried into those areas from upstream locations. However, it is also possible that molluscs once inhabited those areas but have died.
As in Indiana Harbor, the only molluscs in the grab samples from the Saginaw River were fingernail clams (Sphaeridae) (Tables 7-13 and 7-14). Musculium sp., Pisidium sp, and Sphaerium sp. were collected during the December 1989 survey, but only Musculium sp. was present in the June 1990 survey. The reason for the difference in species composition between seasons is unknown. Perhaps these genera were present at very low abundances during the June 1990 survey and were missed when the samples were collected, or perhaps they were not present at any of the stations during the June 1990 survey because of changes in available habitat or increased contamination in these areas.
Comparisons among the three AOCs indicated that Indiana Harbor (Station 10 excluded) may have been the most toxic AOC for benthic invertebrates, and the Buffalo River may have been the least toxic AOC (Table 7-15). This conclusion is based on the overall numerical abundances (BR>SR>IH) and total number of species (BR=33>SR=20>IH=14) present at each AOC. The Buffalo River had many genera that were not present or were present in low numbers in Indiana Harbor and Saginaw River. Varying degrees of contamination at these AOCs could influence the abundance and species composition of benthic communities, but the influence of differences in substrate quality (particle size) and the amount and quality of food sources cannot be discounted. Regression analysis indicated no significant relationship between benthic communities and any single contaminant or physical variable at any of the stations. Given the complex mixtures of both organic and inorganic contaminants found at most of the Great Lakes AOCs, it is not surprising that such simple relationships are difficult to discern. A large bed of submerged aquatic vegetation at Indiana Harbor Station 10 provided structural support above the contaminated sediments and most likely a place for relatively clean organic material and associated bacteria to accumulate. As a result, extremely high abundances of oligochaetes were found at Indiana Harbor Station 10.
Percent contribution of major taxon, abundance, and species composition proved to be good discriminators of benthic community responses within and among AOCs. By examining the percent contribution of each taxon to the overall community, it is readily apparent that among the three AOCs, Indiana Harbor was dominated by oligochaetes with very few other benthic taxa present. This result is also apparent from the numerical abundances and the species composition of the various taxa of invertebrates. All three of these measures are easily obtained, and statistical comparisons among numerical abundance estimates are readily comparable. Overall, the benthic invertebrate communities showed a graded response from low total numerical abundance in Indiana Harbor (Station 10 excluded) to high total numerical abundance in the Buffalo River. Results of sediment chemistry analyses and laboratory toxicity tests indicated that Indiana Harbor was the most toxic AOC, while Saginaw River was the least contaminated. The three measures of sediment contamination (i.e., benthic invertebrate community, laboratory sediment toxicity testing, and sediment chemistry) do not always agree, and the reasons for these discrepancies need to be identified. However, the integrated sediment assessment approach will help reduce the Type I and Type II errors when assessing contaminated sites.
Comparisons between the concentrations of simultaneously extracted metals, total PAHs, and total PCBs and invertebrate abundances demonstrated a consistent pattern of decreasing abundance with increasing contamination (Figures 7-2, 7-3, and 7-4). Regardless of the contaminant examined, there seemed to be a concentration below which abundance was independent of contaminant concentrations. There also was generally a concentration above which abundance was consistently reduced. This suggests that a threshold concentration of contamination exists, below which invertebrate abundance is more strongly controlled by other factors and above which the influence of the contaminant is more pronounced. Therefore, additional research is needed to evaluate specific contaminant, biotic, and abiotic factors that control invertebrate abundance and benthic community structure in contaminated sediments.
The value of this threshold level, regardless of the contaminant examined, seems to be consistently higher for the oligochaetes compared with other benthic invertebrates. This is consistent with the identification of oligochaetes as among the most contaminant-tolerant of benthic invertebrates (Hilsenhoff 1982, 1987; Bode 1988). Relationships between benthic invertebrates and sediment chemistry might have been stronger if the sediment samples for chemistry and physical measurements had corresponded more closely with the benthic invertebrate samples. The benthic grab samples were collected at each station before sediment samples were collected for analytical chemistry, physical characterization, and toxicity testing. Benthic invertebrates often exhibit patchy distributions (Elliott 1977) and typically have a high variability associated with corresponding abundance estimates (Winner et al. 1980; Luoma and Carter 1992). Sediment physical and chemical characteristics may also be variable (Burton 1991). By not pairing each benthic sample with the chemical and physical samples, the chance for conflicting results was increased.
Deformities in Chironomids
Chironomid genera exhibit different tolerances to sediment contaminants (Hamilton and Saether 1971; Hare and Carter 1976; Warwick 1985, 1988; Wiederholm 1984). Some genera are intolerant, and low contaminant concentrations eliminate them from the benthic community. On the other hand, genera such as Procladius, Chironomus, and Cryptochironomus, are more tolerant (Warwick 1985; Bode 1988). A relationship between increased sediment contamination and the presence of deformities in chironomid larvae has been documented by many investigators (Hamilton and Saether 1971; Warwick 1980, 1985; Tennessen and Gottfried 1983; Cushman 1984; Wiederholm 1984). Some of the reported deformities are thickening of the exoskeleton, enlargement and darkening of the head capsule, asymmetry in mouthparts, missing or fused lateral teeth, and deformed antennae.
The mentum (Orthocladinae and Chironominae) and ligula (Tanypodinae) of chironomid larvae were examined for deformities at all three AOCs. The specimens had various mouthpart deformities, including missing lateral and central teeth, asymmetry in the mentum, badly deformed and twisted lateral teeth on the mentum, and missing teeth on the ligula (Procladius). None of the specimens exhibited deformed antennae. Most deformities in this study were found among larvae of the genera Procladius and Chironomus. Even when other chironomid genera were present, they rarely displayed mouthpart deformities. The reasons for this are not clear, but may be the result of individuals dying before they can exhibit abnormalities.
In unimpacted areas, the prevalence of deformities in chironomids is generally less than 1 percent (Wiederholm 1984; Warwick et al. 1987). Several investigators have suggested that deformity prevalences of 5-25 percent or greater are indicative of moderate to severe sediment contamination (Wiederholm 1984; Warwick et al. 1987). Given this criterion, chironomid deformity prevalences at the three AOCs were in the ranges found for moderately to severely contaminated environments (Table 7-16). The stations at which no chironomids were found with deformities were also the areas at which few or no chironomids were collected. Only two individual chironomids were collected from Indiana Harbor (Station 10) and both had deformities.
As with the abundance data, the prevalence of mouthpart deformities in the Buffalo River chironomids indicates that conditions there are less toxic compared with conditions in Indiana Harbor or the Saginaw River. The prevalence of mouthpart deformities was consistently high at the Saginaw River stations, indicating that contaminant concentrations were high enough to affect the chironomid community at most stations. This would seem contrary to the expected result based on laboratory sediment toxicity tests and sediment chemical analyses, which indicate that the Saginaw River samples were less toxic than sediment samples from Indiana Harbor or the Buffalo River. It is possible that the contaminants in the Saginaw River were present in a more available form than they were in the Buffalo River, thereby causing more mutagenic effects than were observed in the Buffalo River.
The occurrence of deformities among chironomids at the three AOCs may be related to the degree of sediment contamination. However, the limited number of samples made it difficult to evaluate these potential relationships. The prevalences of deformities exhibited by chironomids exposed in laboratory sediment toxicity tests (Nelson et al. 1993) should be compared with the prevalences of deformities in chironomids collected from the field. A more specific study is needed, which is designed to elucidate the relationships between particular contaminants and chironomid mouthpart deformities and to encompass a broad range of contaminated and uncontaminated areas. With a minimum of training, chironomid larvae can be mounted on slides and mouthpart deformities can be noted. Further study is needed to determine the usefulness of chironomid larvae mouthpart deformities for identifying sediment contamination. Preliminary results from the ARCS Program indicate that the use of larval chironomid mouthpart deformities in future sediment assessment studies would be useful.
Artificial Substrates vs. Grab Samples
Several investigators have examined the advantages and disadvantages of using artificial substrate samplers to assess benthic invertebrate communities (e.g., Cairns 1982). The artificial substrate samplers used in the ARCS Program were modified from those used by Stauffer et al. (1976). These samplers were easy to work with, and the synthetic mesh was easily pulled apart, allowing easy access to the invertebrates, with a minimum of specimen damage.
Artificial substrate samplers can be deployed at any site to provide a standardized habitat. In the ARCS Program, artificial substrate samplers were deployed at each site for 30 days, and the invertebrate samples were processed following the methods outlined in the Rapid Bioassessment Protocols (Plafkin et al. 1989). At Indiana Harbor Station 03, each artificial substrate sample contained thousands of newly settled (<500 um) zebra mussels (Dreissena sp.). To assess the presence of other benthic organisms at this station, a 100-organism subsample of invertebrates other than zebra mussels was picked from each artificial substrate sample. The invertebrates were then enumerated into major orders and families.
The artificial substrate samplers used in the ARCS Program collected a broader spectrum of benthic invertebrates than the benthic grab samplers (Figures 7-5, 7-6, and 7-7). This finding indicates that there were more benthic invertebrates in the areas sampled than would be collected in the benthic grab samples alone. The absence of certain benthic invertebrates from the sediments collected with the Ponar grab sampler does not necessarily indicate effects of sediment contaminants. Furthermore, LaPoint and Fairchild (1992) caution that colonization of artificial substrates is a function of habitat availability and may not necessarily reflect sediment exposure. The artificial substrates may simply act as a focal point for colonization by invertebrates in areas where other suitable substrates are unavailable. Planktonic larvae or mobile benthic invertebrate species may be present in the water column and settle on the artificial substrate samplers, but not settle on the sediments.
Alternatively, the difference in faunal composition observed between the artificial substrate samples and the benthic grab samples may provide useful supplementary information for benthic community assessments. Benthic grab samples may not effectively sample all taxa that could potentially influence benthic community structure (i.e., through competition or predation) and may therefore suggest a severely impacted community based on the low diversity of taxa. Additional invertebrate taxa may be present in an area that are not collected in the benthic grab sample. Based on the results of the artificial substrate sampling, this information becomes extremely important if potential food chain bioaccumulation estimates are being considered. Therefore, depending upon study objectives, the use of artificial substrate samplers may be warranted because they collect the epibenthic community more readily available to vertebrate predators. Future studies should consider the use of both benthic grab samplers and artificial substrate samplers to make estimates of the total benthic invertebrate community structure.
Variation in Benthic Invertebrate Sampling Using Ponar Grab Samplers
It is important to identify sources of variability in ecological studies so that a meaningful interpretation of the data can be made (Collins and Sprules 1983) and future studies can be designed to address the variability. The results of the variance partitioning in the ARCS Program indicate that among-station and among-replicate variability accounted for most of the explained variability in the abundance estimates (Table 7-17). It is not uncommon for among-station variability to account for a considerable amount of the explained variability in abundance estimates for invertebrates (Lewis 1978; Threlkeld 1983). The variability among stations may be due to 1) heterogeneity of chemical concentrations in sediments, 2) stations being located at variable distances from contaminant sources, 3) differences in substrate characteristics among stations that could influence colonization by the invertebrates, or 4) different station depths that could influence benthic communities.
The variability among AOCs was relatively high for oligochaetes and bivalves, but minimal for chironomids and gastropods. The overall numerical abundances of benthic invertebrates were similar among the three AOCs. This is not unexpected considering that all three AOCs have received substantial amounts of contaminants from industrial and municipal sources, and the majority of the benthic community consists of oligochaetes. Future assessment studies would likely benefit from the inclusion of a relatively uncontaminated reference area.
The relatively high among-replicate variability in abundance estimates was likely due in part to the patchy spatial distributions that most benthic invertebrates typically exhibit (Elliott 1977). For most studies of benthic invertebrates, this source of variability can be reduced by collecting additional replicate samples at each station.
The partitioning of the variance into different components indicated that future studies might provide better data if 1) sample replication was increased, perhaps in conjunction with the use of a smaller grab sampler, and 2) additional stations were sampled to better represent the entire range of sediment contamination within each AOC. The main reasons for limiting the number of grab samples taken from each AOC are the time and cost of processing the samples. By taking smaller individual samples while still sampling the same overall area, the overall processing time for each station would not change substantially, but the estimates of invertebrate abundances among replicate grab samples should have a lower variance (Frederickson 1992, pers. comm.).
Value of Benthic Community Structure Analyses for Assessing Contaminated Sediments
Analyses of benthic community structure provide important information regarding the in situ effects of contaminants on resident biota. By evaluating community structure and the abundances of the genera and species of benthic invertebrates at a site, an assessment can be made of the extent of contamination. This is a useful tool for conducting reconnaissance surveys to determine if a problem may exist at a site. The structure of the benthic invertebrate community can be examined relatively quickly at a reasonable cost to provide a qualitative assessment of how the benthic invertebrate community may be affected by contaminants. Although benthic invertebrate community studies provide evidence of sediment contamination, they cannot identify the contaminants or even families of contaminants that are responsible for adverse effects.
The structure of benthic invertebrate communities should be evaluated in future assessments of Great Lakes sediments. A tiered approach using benthic invertebrates may be warranted. The first tier should be a qualitative assessment of the benthic invertebrate community to 1) determine if the community structure shows signs of alterations relative to the community structure in unaffected areas; 2) evaluate whether there are differences in benthic invertebrate communities across spatial gradients that may identify potential hot spots of contamination; 3) determine if representatives from several orders are present, and determine if the community is skewed toward one or a few orders such that a second tier is unwarranted; and 4) determine the number of samples required in the second tier of the assessment.
The benefit of using a first-tier assessment to examine the resident benthic invertebrates in a contaminated area is that decisions can then be made on the best methods for sampling, based on the organisms that are present. The first-tier assessment can help determine how detailed the sampling plan should be, or even if benthic sampling should be conducted at all in a second-tier assessment. For example, in the ARCS Program, the majority of the benthic organisms collected were oligochaetes and chironomids. Some of the oligochaetes and chironomids are very small and would pass through a 500-um mesh sieve. A first-tier assessment would have indicated that the community was primarily oligochaetes and chironomids, with very few other benthic invertebrates. Therefore, a 250-um mesh sieve could have been used instead of a 500-um mesh sieve when collecting the invertebrates in a second-tier assessment. Considering the low diversity of benthic invertebrates, the decision may also have been made that a second-tier assessment was unwarranted.
The second-tier assessment, if warranted, should involve a more quantitative analysis, if the qualitative analyses in the first tier warranted further evaluations of the sites. Based on the first-tier assessment, the decision as to whether a second-tier assessment is warranted will be different for each project, considering the objectives of the project and the funding constraints. For instance, in the ARCS Program, oligochaetes and chironomids comprised over 90 percent of the benthic invertebrate community at all of the stations examined. At this point, a decision could be made that the species composition of these two orders is not sufficiently important, so further analysis would not be warranted. However, a decision could also be made that knowledge of the species composition is important, so that the success of any remediation activities can later be evaluated. Quantitative measurements, including statistical analyses of various community measures, would then be appropriate in a second-tier assessment of the benthic invertebrate communities.
Benthic invertebrates can be used to monitor the success of remediation activities. A sampling regime could be designed to monitor the long-term recovery of the benthic invertebrate communities in the remediated areas. This type of monitoring is important because the recovery of benthic invertebrate communities may affect the whole aquatic ecosystem. By comparing the structure of benthic invertebrate communities at remediated sites with the communities at undisturbed reference sites, the rates and effectiveness of recovery can be quantified and monitored.
Oligochaetes and chironomids accounted for more than 90 percent of the benthic invertebrate community collected using the Ponar grab sampler in the three AOCs. The dominance by these two taxonomic groups is indicative of disturbed benthic invertebrate communities. Indiana Harbor was the most toxic AOC for benthic invertebrates. The Buffalo River had the largest number of genera and species (n=33), followed by the Saginaw River (n=20) and Indiana Harbor (n=14). This result contrasts with laboratory toxicity tests and sediment chemistry evaluations, which predicted the Saginaw River as the least contaminated AOC.
Comparisons between concentrations of simultaneously extracted metals, total PAHs, and total PCBs and benthic invertebrate abundances demonstrated a consistent pattern of decreasing abundance with increasing contamination. These data suggest that a threshold concentration exists below which abundance may be controlled by factors other than contaminant concentrations and above which the influence of contaminants is pronounced. However, direct cause-and-effect relationships between invertebrate abundances and individual contaminants were not demonstrated.
The prevalence of larval chironomid mouthpart deformities was relatively high in all three AOCs. The Buffalo River had a lower prevalence of deformities than the Saginaw River. The two individual chironomids collected from Indiana Harbor were deformed. Overall, the prevalences of deformities in the three AOCs indicate that these areas are moderately to severely polluted.
The Ponar grab samples and artificial substrate samples indicated the presence of different numbers and taxa of benthic invertebrates. While the grab samples were predominantly comprised of oligochaetes and chironomids, the artificial substrate samples were comprised predominantly of amphipods, isopods, turbellarians, and zebra mussels (Dreissena sp.). The Ponar grab sampler may not have sampled a considerable number of benthic invertebrates. The incorporation of artificial substrate sampling into benthic invertebrate surveys may enhance the accuracy of estimates of the total benthic invertebrate community composition and the potential for recruitment to uncontaminated sediments.
The results of a variance partitioning analysis indicated that among-station and among-replicate variability accounted for most of the explained variability in the estimates of invertebrate abundances from grab samples. The variability associated with differences among AOCs was relatively high for oligochaete and bivalve abundance estimates, but minimal for chironomids and gastropods.
Based on the information presented in this chapter, the following conclusions and recommendations can be made:
- Benthic community evaluations provide empirical information on the effects of contaminated sediments on resident biota that are not addressed directly by laboratory toxicity tests or sediment chemical analyses. Changes in benthic communities are likely the result of long-term exposures to chemical contaminants and are, therefore, indicative of chronic effects. Benthic community evaluations should be used as part of an integrated assessment of contaminated sediments.
- Measurements of chemical and physical variables should be made on subsamples of the sediments from which the invertebrates are collected to avoid the potential problems associated with heterogeneous distributions of organisms and contaminants.
- If possible, a preliminary survey of each study area should be conducted to identify the resident benthic taxa and to determine the number of invertebrate samples that need to be collected.
- It is useful to compare the prevalence of mouthpart deformities in larval chironomids exposed to contaminated sediments in laboratory toxicity tests with the prevalence of mouthpart deformities observed in larval chironomids collected from the field.
- Comparisons between the concentrations of certain contaminants and invertebrate abundances suggest a threshold concentration of contamination below which invertebrate abundance is more strongly controlled by other factors and above which the influence of the contaminant is more pronounced. Therefore, additional research is needed to evaluate the specific contaminant, biotic, and abiotic factors that control invertebrate abundance and community structure in contaminated sediments.
- Benthic community assessments should consider using both artificial substrate samplers and grab samplers.
- Because most of the variance in abundance estimates of benthic invertebrates appears to be associated with differences among stations and among replicates, future studies should sample more stations and collect more replicate samples, perhaps using a smaller grab sampler.