DECIDING WHEN TO INTERVENE

Presentation modified from the report:

Moran, T., Adams, D., Walker, K. and B. Zajdlik. 1997. Lambton Industrial Society 1994/1995 St. Clair River Sediment Program Defining Spatial Extent and Environmental Conditions. Prepared for the Lambton Industrial Society by Pollutech EnviroQuatics Limited, Point Edward, Ontario.

Introduction

Results of an integrated sediment study conducted by the Lambton Industrial Society (LIS) in 1994 and 1995 demonstrated continuing concerns with contaminated sediment and associated effects on organisms exposed to the sediment. The project hypothesis "Contaminated sediment is causing deleterious impacts on the aquatic biota of the St. Clair River" was accepted for specific locations in each of the three study zones.

The benthic community structure at 17 of 28 sample sites could not be differentiated from reference sites upstream and downstream of three study zones. Nearshore sites in Zone 3 were classified as moderately impaired, and specific sites in Zones 1 and 2 were slightly impaired. No sample locations were found to be degraded.

The study established an increase in the number of taxa of benthic macroinvertebrates relative to earlier studies at every site, an indication of continuing recovery of sediment quality.

Rationale

Extensive monitoring of the St. Clair River has been ongoing since the mid 1950s. Early studies found a degraded environment characterized by poor water and sediment quality. Subsequent studies indicate a trend of continuing improvement, attributed to reductions in contaminant loadings from industries and municipalities located along the Ontario side of the river. Using the most recent studies, the St. Clair Remedial Action Plan identified the three sediment zones as "priority 1" areas for study due to sediment contamination and impacts on the benthic macroinvertebrate community.

The LIS initiated a study in 1994 to further define the spatial extent of impairment and sediment quality conditions in these three zones. The project hypothesis "Contaminated sediment is causing deleterious impacts on the aquatic biota of the St. Clair River" was adopted.

Methodology

An assessment of the biologically active surficial sediment using an integrated study design was used. The study incorporates bulk sediment chemistry analysis, benthic macroinvertebrate assemblage assessment, and laboratory toxicity tests using a variety of aquatic test species.

Synoptic samples were collected at each of 31 sites using a ponar dredge. Samples were collected at midshore locations in the three study zones in the Spring of 1994, along with nearshore and offshore samples in Zone 3. Additional nearshore and offshore samples were collected in Zone 2 in the Spring of 1995. Reference samples were also obtained from locations near the head of the river, Sarnia Bay immediately upstream of study Zone 1, and downstream of Zone 3 in 1994.

Chemical analyses were completed for compounds for which Ontario has established biological effects criteria, or have been identified as a sediment quality concern by the St. Clair River RAP, or are associated with the prediction of a normal benthic community.

Sediment toxicity tests used in this study included a 21 day fathead minnow survival test, a 21 day mayfly larva survival and growth test, a midge larva 10 day survival and growth test, and an aquatic worm 28 day survival and reproductive success test.

Results and analysis

Field observation. Oil droplets were visible in all Zone 1 and 2 and some Zone 3 samples. Petroleum odors were also evident in these samples. Sewage fungi were found on the surface of all Zone 1 samples. Aquatic plants were present in all samples except for several from Zone 3. Hydrocarbon contamination was observed at reference stations 2 and 3.

Sediment chemistry. Chemical analyses showed that sediment from all stations, including the reference sites, exceeded Lowest Effect Levels (LEL) identified in the Provincial Sediment Quality Guidelines (PSQG) for several compounds. The LEL is considered to be the level of contamination that is tolerable by most benthic organisms. LELs were exceeded at one or more stations for PAHs, individually as well as total PAH, hexachlorobenzene, PCBs, arsenic, cadmium, copper, lead, mercury, nickel, zinc, total organic carbon, and total Kjeldahl nitrogen. Pesticides were below detection limits at all sites.

Hexachlorobenzene at 8 stations, and mercury at 11 stations exceeded the PSQG Severe Effect Levels (SEL). The SEL is the concentration at which a pronounced disturbance of the benthic organisms could be expected. Concentrations of hexachlorobenzene and mercury were lower than historical levels at most locations.

Statistical analysis of the physical/chemical data found that variation among the sites could be attributed to three principal components. The first identified organically or nutriently enriched stations in areas of reduced current velocity and increased deposition. The second identified stations with elevated levels of chlorinated organics, PCBs, chromium, mercury, and zinc, all in Zone 1. The third identified an offshore station in Zone 2 where the highest PAH concentrations were recorded.

Laboratory toxicity testing. Sediment from 27 of the 31 sample sites demonstrated toxicity to at least one species. No location demonstrated toxicity to all four test species. All reference sites showed a toxic (acute or sub-lethal) response in one test species, a typical result of sediment toxicity observed even in pristine areas of the Great Lakes.

Fathead minnows demonstrated the greatest acute toxicity to sediment from many stations within Zones 1 and 2. A partial acute response by mayfly larvae to sediment from two Zone 3 stations, and to midge larvae exposed to one Zone 3 sediment also occurred.

The response by fathead minnows, given their ecological niche, suggests that contaminants are leaching from the sediment into the water during the static toxicity test. This requires further study, as the test condition differs markedly from in situ conditions where the water is constantly refreshed.

The worm test species, Tubifex, demonstrated no acute toxicity to sediment from any test site, but had the highest number of chronic responses based on reduced reproduction. This increased sensitivity is likely related to the intimate contact with and ingestion of sediment-associated anthropogenic toxicants. Although the worms were the most sensitive species to test sediment in laboratory conditions, the in situ benthic community at several of the same stations was dominated by worms. It is likely that an adaptation to in-place contaminants has occurred in the resident worm populations, many of which are not T. tubifex.

Efforts to relate sites statistically on the basis of toxicity responses alone separated the fathead minnow results from those of the three benthic organisms. Sites at which samples of sediment were toxic to fathead minnows tended not to be toxic to benthic organisms, and where toxicity to benthic organisms was evident, fathead minnows showed no response.

Benthic community assessment. The benthic macro-invertebrate community assessment showed an increase in the number of taxa present for all stations within all three zones, compared to studies completed in 1985 and 1990.

Midge and worm taxa together accounted for 73.2-98.0% of the organisms present. Worms dominated at most of the locations described as degraded or "priority 1" in previous studies.

Two well recognized statistical approached to describing variability among stations based on the benthic macroinvertebrate assemblage were applied to the study results.

Principal component analysis (PCA) identified very little variation, separating only two Zone 3 nearshore stations into a grouping distinct form the others, due to the numerical dominance of immature worms at these stations. Reference stations were included in the general grouping.

The lack of variability identified by PCA analysis may be due to a number of factors, including:

• An inadequate number of stations to differentiate the physical/chemical features of the study zones
• Changes in benthic community may be small, requiring a larger sample size to identify variability
• The stations represent a heterogenous environment, as indicated by the wide scatter of the station scores
• No difference exists between the three study zones and the reference stations
• An inability to determine an effect using benthic community assemblage

An alternative statistical examination of results did identify variability. Cluster analysis identified the likely presence of four distinct clusters of sites. Two clusters could be distinguished from the reference sites. The first included stations from Zones 1 and 2 that were also identified in statistical analysis of physical/chemical parameters as being organically rich depositional zones. The second included stations within Zone 3 numerically dominated by worms. Stations in these two clusters include the stations identified in the 1990 study as degraded, as well as additional stations in close proximity to those stations.

All other stations, which were indistinguishable from reference stations, had a reduced number of worms, relatively more midge larvae, and a higher representation of more pollution sensitive species.

Classification of degree of impairment of an area based on benthic assemblage is subjective; however, some conclusions have been drawn. Based on classifications used in the 1985 and 1990 studies, the cluster of stations within Zone 3 that were distinguished from the reference stations remain moderately impaired. All other stations, including sites within each of the three study zones, were indistinguishable from the reference areas. The benthic assemblage in these locations is similar to benthic assemblages outside the study zones.

If it is assumed that the reference stations adequately reflect the benthic assemblage that would be present within the three study zones had chemical contamination not been a concern, then the stations that differentiate from the reference stations likely do so due to chemical contamination within the sediment or in the water column.

Definitively differentiating benthic communities would require incorporating an increased number of reference stations to improve confidence in defining a typical benthic assemblage for areas not affected by chemical contamination. Adapting an approach by Reynoldson (Reynoldson et al. 1995) to assessment of sediment in the open waters of lakes could be accomplished by incorporating reference sites from the U.S. side of the river. Work by Harris (1996) suggested an increasing number of taxa with downstream distance, on both sides of the river.

Integration of results

A fully integrated study requires that the three components of the study be interpreted simultaneously. Chapman et al., (1992) provides an approach for integrating the chemical analyses, laboratory toxicity results, and benthic community assemblages.

The following conclusions are drawn following this approach:

• Contamination is having an effect at all study sites, including the reference sites.
• At four sites, the interpretations "contamination is not biologically available" or "alteration is not due to toxic chemicals" apply. While LEL or SEL levels were exceeded at these sites, no responses were identified in laboratory toxicity tests, and the benthic community was not altered.
• At 20 sites, "chemicals are stressing the system" - some chemicals exceeded LEL or SEL values, and there was a limited toxicological response.
• At 7 sites, there was "strong evidence of pollution induced degradation from sediment".

The conclusion that "chemicals are stressing the system" at the reference sites is based on LEL exceedances and limited toxicity responses, suggesting that these stations are inappropriate as reference stations, or that the interpretation is conservative. Some workers have found that there are variable toxicity responses in areas within the Great Lakes considered to be healthy or pristine. Bailey et al. (1995) concluded that "there is a range of unimpaired communities with associated levels of toxicity, rather than a single, defined, healthy ecosystem." Krantzberg (1994) states that there must be an understanding of the "expected degree of variation in natural communities." A wider selection of reference sites in the St. Clair River is required to resolve these issues.

Interrelationships. A number of statistical approaches were applied to determine overall relationships between and among the three study components.

A Mantel's test comparison of the three data matrices was completed. Mantel's test indicated that each component of the integrated study was independent of the others. No relationship was established between chemical analyses and either sediment toxicity results or benthic community structure, nor between sediment toxicity results and benthic community structure. Benthic community structure and laboratory toxicity results were not related to chemical analyses, nor were toxicity results and benthic community structure related.

Attempts to relate toxicity responses to physical/chemical parameters statistically did not produce robust associations, tending to reflect the naturally occurring physical and chemical parameters rather than anthropogenic parameters. Several contaminants showed weak associations with increases in toxicity, including aluminium, cadmium, copper, zinc, 2,4,5-trichlorobenzene, hexachlorobenzene, octachlorostyrene, and pentachlorobenzene. Others showed no detrimental association with toxicity for any test species - PCBs, total PAHs, and oil and grease.

Presentation of results

Three methods of summarizing integrated results of the study were used.

Detailed summary tables were prepared to interpret data for each site, ie., "chemicals are stressing the system" using the format discussed by Chapman et al. (1992).

Data were also presented visually, linking results at each sample location to its mapped location. Each site was represented by a color-coded pie chart in which the three study elements - chemistry, toxicity, and benthic community structure - are given equal areas. The toxicity response area was subdivided to allow representation of results for the four test species. Three colors were used to represent degree of concern. The visual representation loses some information in presentation, particularly for chemical results, in not distinguishing between one parameter exceedance and many exceedances.

A numerical sediment quality index was derived from the visual presentation as a further simplification. The index weights each of the three study elements evenly, and converts them to a normalized score with a maximum value of 100. Based on the reference stations, an index above 80 would represent normal or unimpaired conditions.

Summary

Overall summary tables, graphical depictions, or numerical indices all attempt to simplify detailed environmental study results into digestible interpretations. In the final interpretation of sediment quality, monitoring results should be based on the full range of detailed information available that has been derived from the application of several accepted assessment tools. This interpretation should also include an understanding of the study area's history with regard to environmental quality, level of contaminant source control, as well as occurrence of historical remediation efforts and resulting implications on environmental quality.

References

Baily, R. C., Day, K. E., Norris, R. H. and T.B. Reynoldson. 1995. "Macroinvertebrate Community Structure and Sediment Bioassay Results from Nearshore Areas of North American Great Lakes." Journal of Great Lakes Res. 21(1):42-52.

Chapman, P. M., Power, E. A. and G. A. Burton Jr. 1992. "Integrative Assessments in Aquatic Ecosystems." Sediment Toxicity Assessment. G. A. Burton, Jr., (Ed.). Lewis Publishers: Chelsea, MI. pp. 313-340.

Harris, I. W. E. 1996. Invertebrate Populations in Near-shore Sediment of the St. Clair River 1990-1995. Unpublished Report. 50 pp.

Krantzberg, G. 1994. "Spatial and Temporal Variability in Metal Bioavailability and Toxicity of Sediment from Hamilton Harbour, Lake Ontario." Environ. Toxic. Chem. 13:1685-1698.

Moran, T., Adams, D., Walker, K. and B. Zajdlik. 1997. Lambton Industrial Society 1994/1995 St. Clair River Sediment Program Defining Spatial Extent and Environmental Conditions. Prepared for the Lambton Industrial Society by Pollutech Enviroquatics Limited, Point Edward, Ontario.

Reynoldson, T. B., Bailey, R. C., Day, K. E. and R. H. Norris. 1995. "Biological Guidelines for Freshwater Sediment Based on BEnthic Assessment of SedimenT (the BEAST) Using a Multivariate Approach for Predicting Biological State." Aust. J. Ecology. 20:198-219.