DECIDING WHEN TO INTERVENE

Introduction

The Northern Wood Preservers Inc. site in Thunder Bay Harbour has, under various owners, produced creosoted wood products such as railway ties and telephone poles, as well as treated lumber using pentachlorophenol, for over 50 years. Earlier studies have indicated that creosote residues have accumulated in sediment adjacent to the site, often to levels in excess of the Severe Effect Levels (SEL) of the Provincial Sediment Quality Guidelines (PSQGs) (Beak Consultants, Ltd. 1988; Hayton 1989). In addition, dioxins and furans (primarily heptachloro- and octachloro- dioxins and furans) have been identified in sediment adjacent to the site (Beak Consultants, Ltd. 1988). The plant is on a dock 200 m wide that extends approximately 300 m into the harbour. Seepage from the site is believed to be the source of the contaminants.

The Ontario Ministry of the Environment and Environment Canada undertook a joint investigation in 1995 to determine the extent and degree of sediment contamination using biological tests. This information would be used to determine which area needed to be remediated in accordance with the protocol developed by the Ministry (Jaagumagi and Persaud 1996). The protocol required biological effects testing using multiple endpoints when contaminant levels exceed PSQGs (Persaud et al. 1993).

Methods

In order to determine the extent of contamination for cleanup evaluation, dense sampling of the area based on a grid system was undertaken. Preliminary investigation showed that most of the creosote residues were within 100 m of the site. In order to better delineate the gradation within the 100 m zone and develop a detailed sediment contaminant map of the area, sediment samples were collected at 25 m intervals along a total of 14 transect lines radiating out from the dock. Beyond the 100 m zone, samples were collected 50 m apart to a maximum distance of 500 m. A total of 93 stations were sampled for sediment PAH and TOC.

Surficial sediment samples (top 5 cm) were collected with a standard 9" x 9" (23 x 23 cm) stainless steel Ponar grab sampler. Three replicate samples were taken at each location and the top 5 cm from each replicate were combined and mixed to form a single sample. The samples were homogenized from which sub-samples of sediment were collected into appropriate sample containers for analysis. Samples for PAH (scan of 16 individual compounds) and TOC analysis were collected at 71 sites, while additional analysis for metals, PCBs, organochlorine pesticides, chlorophenols, and chlorobenzenes were undertaken at 30 of the sites, as well as at the two control sites. Sampling for dioxins and furans was only undertaken at selected sites along two transect lines and the control site. Standard Ministry analytical procedures were followed for all chemical analysis. These are described in detail in OMEE (1983).

Biological sampling involved a field and laboratory component: benthic community structure and sediment bioassays. Benthic samples were collected with a Ponar sampler along 4 transect lines as well as the two control sites. Samples were washed in the field to remove the fine debris using a U.S. # 30 mesh sieve. Three replicates were collected at each sampling station and the individual replicates were preserved separately in 10% formalin solution. Samples were subsequently sorted in the laboratory using a dissecting microscope, to separate the organisms from the debris. All three replicates were sorted individually, and from these results a mean value for each major taxonomic group was obtained. Subsequently, of the three replicates, the sample closest to the mean was selected for detailed identification of the organisms present. This involved identification to the generic level, with species identification where possible.

Sediment (top 15-20 cm) for laboratory sediment bioassays was collected with a Ponar sampler along the longest transect line (T-5.5; 13 test stations), transect T-EF (3 test stations), and one control station. Approximately 10 L of composited sediment were collected at each site, placed in polyethylene lined containers, and shipped in refrigerated transport to the Ministry laboratory. Details of the standard test procedure are provided in Bedard et al. (1992).

Results

Visual observations noted that the presence of creosote in sediment decreases with distance from the dock along all transect lines. In the area close to the dock (up to 100 m), creosote was often encountered on the sediment surface, especially along the north facing transects. Along one transect, significant quantities of creosote were encountered within 50 m of the dock. In some of these locations (within 25 m), liquid creosote formed over 50% of the sediment sample. Sediment creosote content decreased with distance from the dock. Beyond 100 m, creosote was encountered only as small blobs or drops in the subsurface layers of the sediment. Sediment type along all transect lines was similar, and consisted of a thin layer of fine silt overlying a silt/clay mix.

Chemical analysis. The distribution of PAH compounds in sediment showed that along the north and east sides of the site, sediment is characterized by high concentrations of PAH (up to 16,327 mg/kg), but these decrease rapidly with distance from the dock. Sediment concentrations were typically lower along the southern section of the east side and very low along the south side.

Along the north side, all transects yielded sediment concentrations of total PAH above 300 mg/kg within 25 m of the dock. However, by 50 m levels at most sites were below 200 mg/kg, and by 100 m concentrations were generally below 100 mg/kg total PAH. The exception was one transect where levels were above 300 mg/kg at 75 m from the dock. By 175 m, most sediment concentrations were below 20 mg/kg total PAH, and continued to decline to near background levels with increasing distance.

Transects to the east generally showed lower concentrations in sediment, with the exception of T-EF. This sediment contained substantial amounts of creosote, which is reflected in the higher sediment total PAH concentrations at these sites (up to 1,697 mg/kg). However, by 75 m, concentrations were below 80 mg/kg, and by 100 m were near 30 mg/kg.

Dioxin and furan analysis was undertaken on a limited number of transects. The predominant dioxin compounds in sediment were the hepta- and octa- chlorodibenzo-p-dioxins and the hepta- and octa- chlorodibenzofurans. The lower chlorinated forms were present at very low concentrations or were not detected. Typically, dioxin concentrations in sediment were higher than furan concentrations, with the octa- dioxin the predominant compound.

The distribution pattern of dioxins and furans around the site was similar to the PAH patterns. Concentrations were highest within 25 m of the dock (up to 360,000 pg/g OCDD) and decreased rapidly with distance from the dock. At 100 m, concentrations were less than 60,000 pg/g OCDD along the north and east transects.

Total TEQs for the dioxins and furans were also highest close to the dock and decreased rapidly with distance from the dock. Total TEQs were highest at sites within 25 m (up to 1,320 pg/g 2,3,7,8-T4CDD toxic equivalents), and suggests there is significant toxic and bioaccumulation potential associated with this sediment. However, since I-TEQs are based on mammalian toxicity, they may not be directly applicable to sediment. In addition, the availability of highly chlorinated compounds, such as OCDD are usually overestimated on the basis of partitioning coefficients, since molecular size has been suggested as limiting the passage of large molecules across cell membranes (Smith et al. 1988).

Benthic community structure. Benthic communities at the sample sites consisted primarily of oligochaetes and chironomids. Oligochaete density and diversity did not show any relationship with sediment PAH or PCDD/F levels (benthic samples were not collected in the creosote pool). Chironomid density was found to vary with sediment PAH concentrations, though the correlation was weak (r = -.6794; p<0.05). At distances greater than 150 m from the dock, neither showed a response to sediment PAH concentrations, which in this area were typically less than 30 mg/kg.

Laboratory sediment bioassay. Whole-sediment toxicity tests were conducted using the mayfly nymph, Hexagenia limbata (21-day exposure, survival and growth); the midge larva, Chironomus tentans (10-day exposure, survival and growth); and the juvenile fathead minnow, Pimephales promelas (21-day exposure, survival and chemical bioaccumulation). The battery of sediment toxicity tests used provide a number of endpoints, using organisms representing different trophic levels in order to measure differences in sediment quality. Spatial differences can be ascertained among test sites, as well as against low level contamination using appropriate control sediment.

Conductivity, pH, total ammonia, un-ionized ammonia and dissolved oxygen parameters were measured in the overlying water periodically during the course of the bioassay. pH ranged from 7.0 to 8.2 and conductivity from 279 to 447 umho/cm. Total ammonia readings in the overlying water were elevated for the majority of the test sediment and the reference sediment in the minnow sediment bioassay. Temperature averaged 20^oC to 21^oC for each bioassay.

Mayfly lethality results showed that within 100 m of the dock mortality was significantly higher at certain test sites relative to both negative and reference control sediment (p<0.0073). Sediment collected from Station T-5.5-75 m and T-EF-25 m was found to be acutely toxic (100% mortality). Observations made within the first 24 hours on these test chambers indicated that all of the animals were on the sediment surface. The mayflies showed minimal activity such as swimming or attempts at burrowing, thereby exhibiting strong avoidance behavior. Mayfly avoidance was also noted at Station T-5.5-25 m during the first four days and significant lethality (50% mortality) occurred by Day 21. Mayfly percent mortality was less than 10% for all control and test sediment beyond 100 m from the dock, with no statistical differences reported between the test sediment relative to either control sediment (Dunnett's t-test, p<0.05). Significant differences in the sub-lethal growth endpoint were measured among sites within a 100 to 150 m distance along T-5.5 (p<0.0001). The data, represented by individual fresh weights, showed a 50% growth reduction. Animals exposed to sediment collected from beyond 175 m attained similar or higher weights as the reference control mayflies.

Chironomid lethality and growth results indicate that within 100 m of the dock, significantly higher lethality was noted for three of the test sediment (p<0.0001). After 10 days, percent mortality ranged from 54% to 100%. Percent mortality for the midge ranged from 0% to 17% for sites beyond the 100 m distance. Control mortality ranged from 15% to 16% and was below the acceptable control mortality criterion of 25%. Sediment which yielded poor organism survival also resulted in lower body weights (p<0.0001). Similar to the mayfly assay, a 50% growth reduction in the midge was reported at Stations T-5.5-100 m, -125 m and -150 m and was significantly lower than those attained for control sediment along with the remaining test sediment (p<0.0001).

Fathead minnow lethality results showed that within the 100 m zone percent mortality among treatments were significantly different (p<0.0001). The most toxic sediment was Station T-5.5-75 m (73% mortality) and Station T-EF-25 m (93% mortality). Fish exposed to Station T-5.5-75 m and T-EF-25 m sediment exhibited a loss of equilibrium with a tendency to swim in a vertical manner within 24 hours after their introduction into the test chambers. Avoidance of the sediment, reduced swimming activity, and lack of sediment disturbance continued for at least four days. Mortality first occurred on Day 16 and continued until Day 21. Beyond the 100 m zone, percent mortality for Station T-5.5-150 m (66% mortality) and T-5.5-175 m (56% mortality) was significantly higher than both control minnow survival values. Minnow mortalities began on Day 14 and continued until Day 21. Sediment avoidance behavior was also noted within the first 48 hours for Station T-5.5-100 m and T-5.5-125 m exposures.

There is an association between the concentrations of PAH compounds measured in the bioassay test sediment and the degree of biological effects. The incidence of significantly higher organism mortality was greater for sediment collected within 100 m of the dock. Acute toxicity to the mayfly and midge was measured along the two transects at distances of 25 m and 75 m, respectively. This sediment had an oily sheen and emanated a strong to moderate odor of a creosote-type compound.

Sediment collected between 100 m and 150 m along transect T-5.5 elicited significantly poorer midge and mayfly growth, relative to the sediment collected at a greater distance. Differences appear to be attributable to sediment total PAH concentrations. The LC₅₀ for the mayfly and midge toxicity tests correspond to a sediment total PAH concentration of 150 mg/kg (based on field surficial sediment data). This value is similar to that reported for the amphipod, Diporeia sp., in a dose-response laboratory experiment using PAH-spiked sediment in a 26 day test. Landrum et al. (1991) found a lethal exposure concentration of 100 mg/kg dry weight for total PAHs and the mode of toxic response was attributed to nonpolar chemical narcosis. The lack of minnow toxicity at Stations T-5.5-100 m and T-5.5-125 m appear to be correlated with fish avoidance to the contaminated sediment. Sediment collected at Station T-5.5-150 m and Station T-5.5-175 m, resulted in significantly higher fish mortality relative to the negative and reference control sediment.

Chemical bioaccumulation concentrations in Pimephales promelas are based on unequal sample sizes due to the loss of animals and insufficient biomass across all treatments. A gradient in PAH accumulation was evident. Minnow tissue PAH concentrations were significantly correlated to the total PAH sediment concentrations (r=0.76; p<0.01). The highest total PAH concentrations in minnow tissues was recorded for station T-5.5-150 m (8,844 ng/g), followed by station T-5.5-125 m (3,953 ng/g). Trace amounts were also detected in minnows exposed to station T-5.5-100 m sediment. Non-detectable amounts were reported for the remaining control and test animals sediment (2,680 ng/g) and were representative of pre-exposure conditions.

The significantly lower chemical accumulation by minnows at station T-5.5-100 m, despite the relatively high sediment total PAH concentration of 213 mg/kg, could be due to the stronger avoidance behavior by the minnows. Reduced feeding and sediment disturbance could have resulted in lower chemical uptake. A similar effect, but to a lesser degree, occurred at station T-5.5-125 m. The relatively low accumulation of PAHs in fathead minnows is a result of the ability of many vertebrates, including fish, to metabolize PAHs and their rapid elimination through the bile, feces and urine (Kennedy and Law 1990). The enzyme system that is principally involved in the biotransformation of PAHs is the cytochrome P-450 mixed function oxidase (MFO) system. All these factors would maintain concentrations in the fish at levels lower than those found in the sediment. However, tissue concentrations remain a valuable measure of PAH relative availability.

Discussion

The Ministry protocol requires that where sediment contaminant concentrations exceed the PSQGs SEL guidelines, additional biological assessment needs to be undertaken. Levels of total PAH in sediment exceeded the SEL for total PAH at a number of sites adjacent to the dock (SELs are based on TOC correction and are site-specific).

The biological tests included both benthic community assessment and laboratory sediment bioassays. The biological testing is designed to determine the severity of the contamination. Benthic community studies determine the in-place effects of the contaminants on the existing organisms. Laboratory bioassays assess the effects of contaminants under controlled static conditions of heightened potential availability through both toxic effects (i.e., lethal and sub-lethal effects, such as growth inhibition) and chemical bioaccumulation.

Benthic community structure. The benthic communities within the 100m zone showed effects that could be attributed to sediment PAH concentrations. In particular, the chironomid community showed reductions in density with higher sediment concentrations of total PAH. Along transects T-5.5 and T-7/9, stations close to the dock (25 m) had significantly fewer chironomids and fewer taxa. Since substrate type and depth was relatively uniform along these two transects, the most likely factor was the increase in sediment total PAH concentrations (chironomid density did show a weak negative correlation with sediment total PAH). A simple regression of density versus sediment total PAH suggests that a 50% reduction in chironomid density would correspond to approximately 150 mg/kg total PAH in sediment.

Benthic community structure analysis indicated that beyond the 100 m zone, the benthic community as a whole did not show any direct effects of high sediment concentrations of PAH. Since much of the PAH is present as discrete blobs or drops of oil, it would be relatively easy for most organisms to avoid these areas. This could account for the lack of response to higher PAH concentrations by many organisms. As noted, the distribution of the chironomid fauna does show a correlation with sediment contaminant levels along the north transect T-5.5, and the north-east transect T-7/9 as far as 150 m from the dock, and suggests that sediment PAH is affecting these organisms. Decreases in sediment total PAH concentrations are matched by increases in density of chironomids. The effects on chironomids suggest that below 30 mg/kg total PAH, there is no noticeable reduction in density.

Laboratory sediment bioassay. Sediment bioassay results indicate that there is an increase in both mortality and growth impairment in the benthic species in the sediment close to the dock. Within the 100 m zone, the sediment bioassay results indicate that sediment within 75 m of the dock along transect T-5.5 and within 25 m of the dock along transect T-EF was acutely toxic to both mayflies and chironomids. Sediment from the 100 m to 150 m distance along transect T-5.5 resulted in mayfly and midge growth impairment. At a distance of 175 m and beyond, both growth and mortality were similar to the control values and there was no detectable difference in effects between the test and control exposures. Sediment concentrations were at or below 30 mg/kg total PAH at these distances.

Therefore, at 30 mg/kg total PAH, there appeared to be no effect on these organisms relative to the control stations. Sediment bioassays tend to augment any impacts of sediment-bound contaminants. The process of preparing the sediment prior to testing results in a more complete mixing of any contaminants throughout the sediment, and also potentially heightens the bioavailability of the compounds through disturbance of the sediment. This test, in effect, simulated expected responses under dynamic conditions where mixing, resuspension, and deposition would occur. As a result, it appears from these test results that sediment up to and including 30 mg/kg total PAH could be left in place with no negative effects on benthic communities.

When the test results for the chironomid and mayfly toxicity tests were plotted against surficial field sediment total PAH concentrations, both the mayfly and the chironomid mortality data indicate that there was an increase in mortality with increasing sediment total PAH concentrations. Greater than 50% mortality was found to occur in sediment from the 25 m to the 75 m distance. This corresponded to the zone where sediment total PAH concentrations were in excess of 150 mg/kg. Regression analysis between surficial field sediment total PAH concentrations and bioassay test results found that the area of 50% mortality of chironomids coincided with the 150 mg/kg concentration of total PAH (lower 95% confidence limit) while the area of 50% mortality of mayflies coincided with approximately 130 mg/kg total PAH (lower 95% confidence limit).

Sediment beyond 75 m showed little toxicity to mayflies and chironomids, but there was growth inhibition associated with the sediment. Only at the 175 m distance, where concentrations in sediment were near 30 mg/kg total PAH, did growth rates increase. Growth rates for both mayflies and chironomids stayed high from 175 m to 500 m and equaled or exceeded levels of the control.

Minnow results indicate that there was an increase in mortality at some stations within 100 m of the dock. Along both transects T-5.5 and T-EF, mortality was highest at those locations where sediment total PAH concentrations were highest (i.e., 25 m along T-EF and 75 m along T-5.5).

The bioaccumulation data showed a gradient of PAH accumulation by fathead minnows such that locations close to the dock resulted in higher tissue residues. By 175 m north from the dock, there were no detectable levels in minnow tissues. Analysis of the data showed a significant correlation between tissue residues and sediment total PAH.

Cleanup strategy. The different levels of biological effects were used to define three zones of contamination. Each would merit a different cleanup strategy.

The first, representing the most contaminated conditions, was the area of heavy, visible contamination of sediment by creosote (a creosote 'pool'). This area was located along transect line T-5 and was found to include the 50 m distance, but did not extend to the 75 m distance. Since transects on either side (T-4.5 and T-5.5) of T-5 did not yield similar quantities of creosote in the sediment samples, this area appears to be confined to less than 50 m on either side. Cleanup of this area should proceed based on visual observation of creosote on the sediment surface. This area represents a continual source of creosote (and PAH contamination) to both the water column and adjacent sediment.

The second zone was defined on the basis of acute biological effects, i.e., greater than and including 50% mortality in the test organisms, and coincides with the area of high PAH (>150 mg/kg) and dioxin/furan contamination (>200 ppt total TEQ). This area should be isolated since the toxic potential of the sediment is very high. The approximate boundary of this zone is the area enclosed within the 150 mg/kg total PAH isopleth.

The third zone can be defined on the basis of sub-lethal biological effects and coincides with the sediment area exceeding 30 mg/kg of total PAH. This area is the area enclosed within the 30 mg/kg total PAH isopleth, and represents the area where contaminated sediment should be confined in order to minimize contaminant effects on aquatic biota. Below this concentration, there was no measurable effect on benthic organisms.

Both contaminant concentrations and biological effects are low or not apparent in those areas below 30 mg/kg, and this area would be suitable for natural remediation since existing contaminant concentrations pose little threat to biota. Comparison with an earlier study by Beak (1988) indicate that surficial sediment concentrations of total PAH have decreased since 1987, likely through deposition of cleaner material on the surface. Active deposition of new material would serve to effectively isolate the relatively more contaminated sediment in the deeper layer and would permit longer term degradation of contaminants in this area with little concern regarding potential exposure to aquatic organisms.

Conclusion

Based on the study results, a site remediation plan was developed in conjunction with the property owners. The plan calls for enclosure of the dock behind a clay barrier since seepage from the site is considered to be the source of the contamination. Outside of the clay barrier the plan calls for construction of a rock berm that encloses all of the area where sediment concentrations exceeded 150 mg/kg total PAH. Clean fill is to be placed behind this structure and is to be brought up to grade level (i.e., dry capped). The enclosed area will also contain a treatment cell that can accommodate 20,000 m³ of sediment which is to be removed from the creosote pool and all areas where existing concentrations of total PAH are in excess of 260 mg/kg. This value is based on Ontario's soil cleanup guidelines for PAH. Soil cleanup guidelines were used since the area to be confined behind the berm will become land. At present, the plans call for biological treatment within the cell. Areas where sediment concentrations of total PAH were below 30 mg/kg would be left to remediate naturally.

References

Beak Consultants, Ltd. 1988. Lake Sediment Studies - Thunder Bay, Lake Superior: Northern Wood Preservers Sediment Sampling Program, 1988. Report to Ontario Ministry of Environment.

Bedard, D., Hayton, A. and D. Persaud. 1992. OMOE Laboratory Sediment Biological Testing Protocol. Ontario Ministry of the Environment. Toronto, Ontario. 23 pp.

Hayton, A. 1989. Report on Sediment Toxicity and Contaminant Bioaccumulation near Northern Wood Preservers Inc., Thunder Bay. Ontario Ministry of Environment.

Jaagumagi, R. and D. Persaud. 1996. An Integrated Approach to the Evaluation and Management of Contaminated Sediment. Ontario Ministry of Environment.

Kennedy, C. J. and F. C. P Law. 1990. "Toxicokinetics of Selected Polycyclic Aromatic Hydrocarbons in Rainbow Trout following Different Routes of Exposure." Environmental Toxicology Chemistry. 9:133-139.

Landrum, P. F., Eadie, B. J. and W. R. Faust. 1991. "Toxicokinetics and Toxicity of a Mixture of Sediment-Associated Polycyclic Aromatic Hydrocarbons to the Amphipod Diporeia." Environmental Toxicology Chemistry. 10:35-46.

Ontario Ministry of the Environment and Energy. 1983. Handbook of Analytical Methods for Environmental Samples. Vol I and II. Ontario Ministry of Environment and Energy. Toronto, Ontario.

Persaud, D., Jaagumagi, R. and A. Hayton. 1993. Guidelines for the Protection and Management of Aquatic Sediment Quality in Ontario. Ontario Ministry of the Environment. Toronto, Ontario. 30 pp.

Smith, J. A., Witkowski, P. J. and T. V. Fusillo. 1988. "Manmade Organic Compounds in the Surface Waters of the United States - A Review of Current Understanding." U.S. Geological Survey Circ. 1007. 92pp.