Home Projects About Links EnviroTools Glossary Search
 


Review of Ecological Risk Assessment - Ashland Lakefront Site, Ashland, WI

By:
Christopher Marwood, Ph.D.
Post-doctoral Research Fellow
Department of Zoology
Miami University
Oxford, Ohio, U.S.A.

For:
Technical Outreach Services for Communities (TOSC)
Michigan State University.

Conducted at the request of Ashland/Bayfield County League of Women Voters and the Lake Superior Alliance

Date: May 14, 2001

Analysis

During this phase of the ERA, stressor exposure and effects are quantified using measurement endpoints, which are quantifiable values related to the assessment endpoints. According to the EPA Guidelines these can include measures of exposure, effect and ecosystem characteristics. These endpoints are quantitatively compared to characterize both the exposure of organisms to the contaminant and the ecological effects.

SEH chose a number of measurement endpoints to assess exposure and effects on benthic organisms (benthic community survey, toxicity of sediment to benthos in laboratory experiments), zooplankton (laboratory sediment exposures) and fish (laboratory sediment exposures and PAH tissue concentrations). Because D&M chose to use the SEH sediment chemistry data, and took no additional measurements, its measurement endpoints were necessarily restricted to those identified by SEH. The striking discrepancies between the SEH and D&M documents in the analysis section of the assessments are due to the differential use of this data (D&M did not use results of laboratory experiments or benthic surveys in its ERA), and different techniques for comparing exposure and effects.

Exposure characterization

SEH characterized exposure by comparing sediment PAH concentrations to sediment quality benchmarks empirically derived from 28-day toxicity tests using Hyalella azteca (HA-28). The database was constructed using data from several contaminated sites in the Great Lakes, as well as other contaminated sites (Ingersoll et al. 1996). These benchmarks represent the concentrations below which toxicity is rarely observed (Effects Range Low, ERL) or above which toxicity is frequently observed (Effects Range Median, ERM). In many of the sediment samples taken during the spring and winter of 1998, PAH concentrations frequently exceeded the ERL, and often the ERM. SEH interpreted these data to indicate exposure of sediment-dwelling organisms to PAHs was likely at these sampling locations.

The approach used by SEH is commonly used in ERAs to characterize exposure. Biological effects-based quality criteria combine field data from multiple contaminated sites in an attempt to classify a site, as to the probable effects on the biota, based on the chemical concentrations in sediment or water. The incidence of biological effects associated with a range of chemical concentrations is determined to establish concentration categories in which, for example, no effect, low effect and severe effects are likely (Long et al. 1995). A variety of biological endpoints have been used as database input, including toxicity from contaminated sediments, toxicity in the laboratory to sediments or water (EC50s), equilibrium partitioning, and benthic community structure (Long et al. 1995). With a large enough database, the probability of certain category of effects (none/minor/severe) at a contaminated site can be estimated with some degree of confidence. During the initial screening level assessment, SEH selected the HA-28 benchmarks from a number of potential benchmarks used by various regulatory agencies, based on several criteria including size of the database, number of freshwater data points, etc. The choice of the HA-28 benchmark seems to have been carefully considered.

D&M criticized the sediment quality endpoints chosen by SEH, on the basis that the sediment quality criteria do not protect populations. D&M's argument that SEH based its conclusions on comparison to benchmark concentrations only is erroneous. SEH used the benchmarks only to assess exposure, and incorporated results of several other lines of evidence during its decision-making process. While D&M disagreed with SEH's choice of assessment and measurement endpoints, D&M chose not to take its own measurements. The absence of additional measurements undermines D&M's criticism of the endpoints selected by SEH.

A major difference between the SEH and D&M assessments was the use of different chemistry measurements - SEH used only the smaller set of measurements taken in 1998 for which organic carbon (OC) measurements were taken; D&M used both the 1998 data as well as the 1996 data, assuming an OC of 4%. However, the use of 1996 data versus 1998 data had no practical impact on the assessment, as D&M made numerous calculation errors in its exposure characterization.

In the D&M assessment, exposure was characterized using a model designed for sediments containing multiple PAHs (?PAH model) incorporating equilibrium partitioning, quantitative structure activity relationships (QSAR), with an additive toxic units approach (Swartz et al. 1995). In this model, organic carbon partitioning coefficients (Koc) are used to estimate interstitial water concentrations from the field sediment concentrations, which are then used in a simple QSAR to generate toxic units (TU) for each PAH, where one TU is equivalent to the concentration of a PAH that kills 50% of the organisms. TUs are calculated as the estimated concentration in the pore water divided by the 10-day LC50, which were estimated using a simple QSAR equation based on empirically-derived LC50s for four aquatic test species exposed to three PAHs (fluoranthene, acenaphthene and phenanthrene). The fraction of organisms likely to experience acute toxicity is predicted from the sum of the TUs for each PAH (?TU), assuming additive toxicity. Models using QSARs and equilibrium partitioning to estimate pore water concentrations, such as the one favored by D&M, are useful for assessing sediment toxicity for chemicals such as PAHs with a narcotic mode of action (Di Toro et al. 2000). However, for a useful analysis the model must be applied properly. There were a number of errors in the application of the model.

Instead of calculating TUs for each PAH in the sediment and summing the TUs, D&M modified the model by calculating an "average" LC50, using a mean octanol-water partitioning coefficient (Kow) for each sampling location. D&M pointed out that the model limits the contribution to toxicity from each PAH according to its water solubility. However, in using an average Kow, D&M disregarded this stipulation of the model. As an example, anthracene from sampling location 2300-1500 was measured at 49,000 µg kg-1 in the sediment, which, according to the Swartz equilibrium-partitioning model represents an interstitial water concentration of 84.35 µg L-1. The model estimates an LC50 of 180 µg L-1 for anthracene, and thus, the anthracene TU for this sampling location would be 84.35 / 180 = 0.469. However, the model places an upper limit of TU = 0.25 for anthracene, based on a maximum aqueous solubility of 44.6 µg L-1. By using an average Kow and ignoring limits on solubility built into the model, D&M overestimated the toxicity of sediments with high PAH concentrations.

D&M further deviated from the Swartz model. Swartz specifies a toxic units approach for comparing the PAH concentrations at each sampling location to the expected LC50, but D&M calculated a hazard quotient (HQ) from the total PAH concentration in each sample divided by an EC25. When dealing with mixtures of chemicals, an approach that acknowledges interactions between chemical species, such as the TU approach, is preferred over an HQ, which is used when considering single contaminants. D&M's rationale for using HQs and EC25 rather than TUs was based on the presumption that protection of fish populations can be achieved at PAH sediment concentrations less than the EC25. The choice of the denominator in the HQ seems somewhat arbitrary. The EC25 was derived from a general equation by Suter (1996; this reference was not identified in the ERA), in which population EC25s (for weight of juvenile fish per egg) were extrapolated from LC50s (96-h in fish) by linear regression for use as an index of population effects where no experimentally derived EC25s were available. Suter has stated this technique has the disadvantages of being both "unconventional" and "not conservative" (Suter 1996).

In addition to problems inherent in D&M's approach, the HQ values reported by D&M seem to be incorrect. HQs based on the EC25 should be greater than the HQ based on the LC50, since the EC25 is by definition less than the LC50. It is assumed that this error, as well as the negative values sometimes reported for the HQ (which is not mathematically possible), is the result of modifying the Swartz model to use "average" LC50s for each sample. There were errors in the PAH sediment concentration data, octanol-water partition coefficients, molecular weights, as well as in the model calculations. Those calculation errors render invalid the results of the analysis phase of the D&M assessment. The overall consequence is that the conclusions of D&M regarding Ashland sediments toxicity to organisms are inaccurate.

values that are similar to those derived by SEH using its HA-28 benchmarks. The similarity of the model with the ERL benchmark was predicted by the authors of the ?PAH model (Swartz et al. 1995). For most Ashland sediment samples, ?TU's were greater than 1. At only one sampling location ?TU < 0.186, a value below which the Swartz model predicts PAHs are unlikely to contribute to toxicity. At all sampling locations except five, ?TU > 3.3, indicating acute toxicity will occur with virtual certainty. Based on those criteria, toxicity from PAHs can be expected in most of the sediments offshore from the Ashland Lakefront Property. Only for those sediments on the outer edges of the contaminant distribution would toxicity to benthic invertebrates be expected to be low.

Ecological effects characterization

D&M excluded from its analysis results from both the benthic surveys and the laboratory toxicity tests conducted by SEH. In the analysis and risk characterization sections of the D&M assessment, sediment PAH concentrations (from 1996 versus 1998 in the SEH assessment) were compared to concentrations expected to cause adverse effects, but the benthic surveys and laboratory toxicity experiments were not included in the decision process. In an addendum referred to as "Verification Studies," these experiments were re-examined, criticized, and dismissed as inconclusive. TOSC believes the exclusion of these data from the risk characterization compromises the conclusions of the D&M assessment.

SEH characterized ecological effects using three distinct analyses: sediment toxic units calculated from PAH concentrations and ERMs, benthic community indices, and laboratory sediment exposures. In addition to several locations sampled for PAH concentrations only, four locations were sampled for the benthic community survey and laboratory toxicity experiments. Two locations were sampled within the area of the bay indicated by previous analyses to be contaminated, and two locations outside this area were used as reference locations. Paired sampling locations, consisting of woody and non-woody substrate, were appropriately used in an attempt to minimize effects of co-variates such as the wood chips.

For each PAH identified in Ashland sediments, a toxic unit was calculated by dividing the sediment PAH concentration by the HA-28 ERM. The TUs for each PAH were then summed, to provide an overall estimated PAH toxicity of the sample. For sediment samples in which the summed TU was greater than one, adverse effects were expected. SEH calculated TUs based on PAH concentrations normalized to dry weight, and to organic carbon. The TU concept is commonly used with complex mixtures of contaminants with similar modes of action, and is similar to the approach used by D&M. The SEH TUs, in contrast with the D&M TUs, appear to be calculated correctly and applied appropriately. SEH found ?TUs greater than one for both the PAH-contaminated sampling locations and the reference wood sampling location.

SEH conducted a survey to compare benthic organisms at each station. A cursory examination of the raw data (counts) suggests a strong correlation between PAH concentration and taxa abundance, with fewer organisms and less diversity in the contaminated samples. SEH calculated several indices related to abundance and species richness. Not surprisingly, there was a strong correlation between sediment PAH concentration and loss of abundance and richness. There was little effect of substrate; abundance and richness were predominantly related to dry weight PAH concentration. These data provide strong evidence supporting SEH's conclusion of detrimental effects of PAHs on benthic communities.

Although the benthic survey was not used in the D&M risk characterization, D&M provided its own interpretation of the results. It was suggested that overlap in the data distribution (due to high variability between the six subsamples at each sampling location) made the results inconclusive. It was also suggested that the presence of the wood chips at the sediment-water interface, rather than PAHs, contributed to differences between samples. None of those arguments to discard the benthic survey data are persuasive. High variability within subsamples is common in field studies and can make interpretation of results challenging. However, variability in the response does not justify discarding data. Field tests should be designed to have sufficient power to detect differences between contaminated and uncontaminated samples. SEH acknowledged this during the planning stage of its ERA, and ensured adequate samples were taken. SEH's plan for collection of samples from both sand and woody substrates was designed to allow the effects from the substrate to be removed. When results were compared for either the wood or sand substrate, there was clearly diminished abundance and richness in contaminated samples in all endpoints except one: total abundance was greater at the contaminated wood sampling location than at the reference wood sampling location. At this location, the abundance was due primarily to high numbers of sowbugs and immature tubificids in some of the subsamples from the contaminated sampling location. In this case, the diminished taxa richness in the contaminated wood samples more accurately reflects the data. Regardless, the variability in the Ashland benthic survey data is not as great as D&M implies, and certainly does not invalidate the results.

The third set of analyses used by SEH to characterize ecological effects were laboratory exposures to sediment using a variety of aquatic species: Hyalella azteca (amphipod), Chironomus tentans (midge), Lumbriculus variegatus (oligochaete), Daphnia magna (cladoceran), and Pimephales promelas (fathead minnow). Using sediment from the four sampling locations in which PAH concentrations were previously characterized (CW, RW, CS, RS), H. azteca, C. tentans and L. variegatus were exposed directly to sediment, and D. magna and P. promelas were exposed to sediment elutriate. Standard test methods for exposure to sediment or elutriate were followed. For all species examined, exposure to sediment from the two contaminated sampling locations reduced survival and growth endpoints. When expressed normalized to organic carbon, there was a decrease in survivorship and weight at higher PAH concentrations. Normalization to organic carbon is required when comparing toxicity from PAHs in sediments, to reduce variability from differences in bioavailability.

D&M had two major criticisms of the laboratory experiments reported in the SEH document. The first criticism, that PAH concentrations were not measured in the sediment samples used in the exposures, appears to be unfounded. Although sediment samples were collected on two dates separated by one week, samples for PAH analysis were taken from the same samples used for the laboratory exposures. The second criticism identified by D&M was that the four locations sampled for both the benthic survey and the laboratory experiments were too few to allow delineation of a no-effect concentration. D&M pointed out that PAH concentrations at the contaminated wood location were much greater than the concentrations at the other three sampling locations, and did not allow an adequate concentration-response curve to be constructed. D&M cited these as major flaws in the SEH study design. While these are valid points, it does not appear that the construction of a concentration response relationship from field-collected samples would be a feasible objective. The heterogeneous distribution of PAHs and woody substrate would result in an unmanageable number of samples and toxicity tests. Although a greater number of sampling locations would certainly increase confidence in the results, the approach used by SEH allowed an adequate description of the toxicity associated with contaminated versus reference sediments.

The SEH assessment also reported a series of UV co-exposure experiments. L. variegatus, D. magna and P. promelas were exposed to sediments under normal laboratory light without UV, and then transferred to clean water under light containing UV. Rapid mortality was observed in organisms previously exposed to the sediment from the contaminated wood sampling location. Two notable conclusions can be drawn from these results. The first is that PAHs were almost certainly the contaminant responsible for toxicity in the Ashland sediments; few other chemicals exhibit photoinduced toxicity. Secondly, the elevated toxicity in UV light compared to toxicity without UV suggests organisms exposed to sediments in the field (i.e., in sunlight) likely experience greater toxicity than indicated by the other laboratory exposures. D&M pointed out phototoxicity is not currently examined in standard assays of most regulatory agencies, and attempted to dismiss phototoxicity as having "tenuous applicability to the field." D&M's characterization of the relevance of phototoxicity contradicts numerous studies that have demonstrated convincingly the importance of evaluating phototoxicity of PAHs in the field (Barron et al. 2000; Marwood et al. 1999; Monson et al. 1995; Tagatz et al. 1983). The absence of phototoxicity from regulatory standards for PAH toxicity assessment is simply a reflection of the recent recognition of impacts in the field.

D&M, in its ERA, suggested phototoxicity in the field may be minimized by the presence of dissolved organic matter (DOM), and attenuation of UV in the water column. Humic acids have been shown in laboratory studies to ameliorate phototoxicity of PAHs dissolved in water (Gensemer et al. 1998; Weinstein and Oris 1999). In bodies of water possessing high levels of DOM, such as small oligotrophic and mesotrophic lakes, DOM may provide some protection from PAHs in the water column. However, as the aqueous phase is only one (probably minor) route of exposure, the attenuation of toxicity of the Ashland sediments by DOM is likely minor. When discussing phototoxicity of PAHs, D&M referred frequently to UVB. It should be noted that PAH phototoxicity also involves UVA wavelengths. As UVA wavelengths constitute a greater fraction of sunlight than UVB, and also penetrate to greater depths in the water column, the induction of phototoxicity in the field from UVA wavelengths is greater than from UVB. Therefore, D&M's statements regarding the attenuation of UVB in the water column by suspended particles have little relevance to the question of phototoxicity. No underwater measurements of light attenuation were taken during sampling. Although water column characteristics vary greatly, typical attenuation coefficients measured at other locations in the Great Lakes (Smith et al. 1999; Williamson et al. 1996) indicate UV levels are likely similar to those used in the laboratory exposures. It is impossible to exactly duplicate all environmental conditions in a laboratory exposure. However, by asserting that the laboratory exposures have no relevance to the Ashland site, D&M has dismissed without justification evidence indicating the possibility of severe phototoxicity to organisms.

Risk Characterization

In the risk characterization phase, exposure and ecological effects are integrated to allow a statement regarding the risk and uncertainty associated with the stressor. Because D&M did not use the results of the benthic survey or the laboratory exposures, its treatment of the risk characterization phase was restricted to estimation of hazard quotients by dividing the sediment PAH concentrations by the EC25s from the Swartz model. D&M characterized the risk to benthic populations as "moderate" in some parts of the contaminated area. However, because of multiple errors in the analysis phase (calculation of EC25s), D&M's risk characterization must be considered flawed as well. Summed TUs correctly calculated as specified by the Swartz model result in an estimation of risk very similar to the SEH risk characterization using the HA-28 SECs.

D&M erroneously stated in its conclusions that SEH based its decision on a comparison of sediment PAH concentrations with benchmark concentrations. While the screening level assessment relied on this data alone to trigger a baseline risk assessment, SEH for its baseline assessment characterized risk using a "sediment triad" approach, in which three separate lines of evidence were examined: sediment benchmarks, benthic community surveys and laboratory toxicity tests. Combined, the corroboration of three different lines of evidence strongly supports SEH's conclusion of "high likelihood of risk" to benthic organisms and juvenile fish. This characterization was based on concentrations of PAHs at many sampling locations in the bay that exceeded the TUs developed from HA-28 ERMs (as well as numerous other sediment effects benchmarks), compared with toxicity observed in the laboratory experiments.

One criticism by D&M of SEH's risk characterization was that SEH's threshold effects concentrations triggering clean up were overly conservative. D&M claimed that on the basis of the toxic units derived from the HA-28 ERMs, ecological effects might be expected to be greater than those observed in the benthic survey and the laboratory exposures. SEH acknowledged in its ERA the HA-28 benchmarks are more conservative than other benchmarks. SEH calculated TUs of 7.1 for the reference sand sampling location, 14.4 for the reference wood sampling location, 119 for the contaminated sand sampling location and 3728 for the contaminated wood sampling location. According to SEH, one toxic unit calculated from HA-28 benchmark concentrations is expected to cause 50% mortality. The toxicity observed in laboratory exposures using other organisms does not support the HA-28 TUs, and therefore, conclusions regarding the toxicity of PAHs based on the H. azteca TU values alone would indeed be conservative. However, SEH compared its TUs with observed toxicity in a dilution study with P. promelas and determined toxicity occurred at organic carbon normalized TUs above 7, and a 20% impact could be expected between 7 and 15 TUs. The observed impacts on benthic populations and organisms in laboratory studies at each of the four sampling locations seemed to follow the TUs calculated by SEH, in that no effects were found with sediments from the sampling location with a TU of 7, minor impacts were observed at the sampling location with a TU of 14, and severe effects were found at the sampling locations with the highest TUs. Because threshold effect concentrations were calculated using results of the benthic survey and all the laboratory exposures, these values are probably not overly conservative, and likely reflect concentrations above which impacts to benthic populations can be expected.

D&M also pointed out that due to the great difference in sediment PAH concentrations between the highly contaminated wood sampling location and the less contaminated locations, an adequate dose response could not be constructed. Obviously, a greater number of sediment samples, spanning a wide range of PAH concentrations, would reduce the uncertainty associated with SEH's threshold concentrations. However, in combination with effects from the benthic survey and laboratory experiments, there is sufficient evidence that the values calculated by SEH represent reasonable sediment threshold concentrations.

Causality

Causality is established by linking the stressor with observed effects. In ERAs driven by observed fish kills, etc, establishing causality is paramount. However, in this ERA the stressor was identified with considerable confidence, and a feasible exposure route from sediment to biota was established. The EPA Guidelines identifies several criteria modified from Koch's Postulates that can be used to establish causality in an ERA:

  1. the effect of the toxicant must be regularly associated with exposure to the toxicant;
  2. effects must be observed when organisms are exposed to the toxicant under controlled conditions (i.e., laboratory conditions);
  3. the same effects found in the laboratory must be observed in the field;
  4. indicators of exposure to the toxicant must be found in the affected organisms.

Although each criterion enhances confidence in the conclusions, not all criteria must be satisfied to reasonably establish causality in an ERA. The results of biological surveys can sometimes be inconclusive because populations sampled may be resistant to the contaminant, resulting in different results from lab tests and field surveys. The presence in highly contaminated sediments of high numbers but few species of benthic organisms suggests this may be the case with the Ashland sediments. The SEH assessment satisfied all except the final criterion. SEH intended to assess PAH residues in benthic organisms collected during the benthic survey, but low numbers of organisms forced this part of the assessment to be abandoned. This is unfortunate, as the presence of PAHs in tissues of benthic organisms would positively verify the source of toxicity in Ashland sediments. Nevertheless, it can be concluded with reasonable confidence from other lines of evidence developed in the SEH assessment, especially the observation of phototoxicity, that PAHs are the source of toxicity.

D&M criticized the "weight of evidence" approach used by SEH to evaluate the likely effects on biota at the Ashland site, suggesting the lack of correspondence between levels of PAHs in sediment and toxicity in some of the laboratory bioassays using this sediment provided evidence contrary to SEH's hypothesis, and required abandonment of SEH's hypothesis that PAHs were causing toxicity. While this approach may be valid for simple testable hypotheses, a more encompassing approach is required for risk assessments. In contrast to laboratory studies, simple statistical tests of null hypotheses (i.e., toxicity or no toxicity) cannot be tested in an ERA, because risk hypotheses are not the same as statistical hypotheses. Because uncertainty is an inherent aspect of risk assessment, an ERA cannot be expected to unequivocally state that adverse effects will, or will not occur in a given ecosystem. Risk assessments in which uncertainty is acknowledged and described are more scientifically defensible and provide risk managers with the information required to make appropriate decisions.

The weight of evidence approach used by SEH, in which causality is supported by various lines of evidence, is more appropriate for field studies (Lowell et al. 2000), and is in fact recommended by the EPA Guidelines during the risk characterization phase (U.S. Environmental Protection Agency 1998). (In the Guidelines, "lines of evidence" is preferred to "weight of evidence," to favor a more inclusive approach, including qualitative evidence). The weight-of-evidence approach is not designed to provide experimental "proof." Rather, it is a process which can be used to assess the likelihood of toxicity. The use of several lines of evidence (positive as well as negative) can increase confidence in the risk decision.

The Guidelines suggests several criteria that should be satisfied when using the lines-of-evidence approach, including the abundance and quality of data, the degree of uncertainty, and the pertinence of the evidence to the original risk assessment questions. In our opinion, the SEH assessment adequately addressed uncertainty and relevance in its risk assessment. D&M considered the benthic survey and laboratory assays invalid due to high variability and lack of conclusive evidence of effects between contaminated and reference sampling locations. While the SEH risk characterization and conclusions might have been reinforced by additional data, specifically with respect to the benthic survey, we consider the D&M characterization of the results as overly critical. Taking into consideration the high variability inherent in field studies, SEH ensured there was sufficient statistical power in its sampling design to evaluate impacts on benthic organisms. While variability was high, there were meaningful differences between the contaminated sampling locations and the reference sampling locations. As part of a triad, lines-of-evidence approach, the benthic survey reinforces other data (lab tests, benchmarks) and enhances confidence in the overall conclusions. To reiterate, we consider D&M's decision to exclude results from the benthic survey and the laboratory exposures as seriously compromising its assessment.

In predicting effects on populations not examined in its ERA, SEH made an extrapolation from the laboratory to the field, and from a small scale to a large one. Extrapolations necessarily increase uncertainty in ERAs, but are unavoidable since direct measurements of impacts are not normally feasible. There are many factors that may alter the response in the field, such as resistance to the contaminant, natural fluxes in environmental parameters, and other physical stressors not present in the laboratory. A major source of uncertainty is interspecies variability in response to the toxicant. SEH chose measures of effect that were closely related to the assessment endpoints, so extrapolation to other trophic levels was minimized. Integrated approaches that use multiple lines of evidence, such as the sediment triad used by SEH, provide the least uncertainty.

Probable Impacts on Organisms

Some inferences can be drawn with respect to the probable current and future effects on biota residing in the Ashland sediments, or otherwise contacting the PAH contaminants.

Nature and intensity of effects

The impact of PAHs in Ashland sediments on populations of benthic organisms may be severe. The concentrations of PAHs measured in the sediments are similar to other contaminated sites in which severe impacts have been observed. Because PAHs tend to remain bound to organic sediments, sediment-dwelling or burrowing organisms are at the greatest risk. The limited benthic population survey conducted for the SEH assessment suggests that benthic communities within the highly contaminated sediments have been impacted by PAHs. Aquatic organisms that reside on the sediment surface, or are restricted to the water column only, are likely to receive less exposure to the sediment-bound PAHs, and are therefore at less risk. However, elutriate exposures conducted in the laboratory suggest aqueous concentrations of PAHs liberated from the Ashland sediments are toxic to aquatic organisms. The demonstration of toxicity to all five species examined reflects the broad range of aquatic organisms susceptible to PAH toxicity.

The SEH risk assessment concluded substantial impacts to populations (a 20% reduction in abundance) are likely to occur when sediment PAH concentrations exceed 7 summed toxic units (normalized to organic carbon). The SEH calculations indicate this level is exceeded in most samples from Ashland. It is reasonable to assume that populations of benthic organisms in a large portion of the offshore area have been negatively affected. In many samples, the summed TUs exceed those calculated for the highly contaminated woody location sampled for the benthic survey and laboratory experiments. Based on the results of the benthic survey, it is likely that the benthic community in most of the affected area is severely diminished. It is reasonable to assume that other aquatic organisms not residing in the sediment have been negatively affected as well, although it is more difficult to measure impacts on these populations because they possess the ability to avoid the contaminated areas.

One possible impact of PAHs identified but not examined in these ERAs is the induction of tumors in fish. Most PAHs have been shown to be mutagenic in laboratory studies. In other PAH-contaminated sites that have been examined in more detail, DNA adducts and elevated incidence of (mainly hepatic) neoplasms has been observed in fish (Ericson et al. 1999; Myers et al. 1987; Vogelbein et al. 1990). Although the contaminated area is small compared to other contaminated sites and the frequency of contact with contaminated sediments is relatively low, mutagenicity and carcinogenicity should not be ruled out as potential impacts of PAHs on fish populations.

Spatial scale

The patchy distribution of PAHs makes it difficult to delineate a precise boundary to the contamination. However, the tendency of PAHs to tightly bind to sediment and remain near to their point of introduction rather than migrating great distances suggests the majority of PAH impacts are likely restricted to the area adjacent to the Ashland waterfront and do not extend beyond the enclosed area identified in the ERAs. The incidence of toxicity and effects on benthic populations are associated with areas in which PAH concentrations are greatest. It is unlikely that severe effects of PAHs extend beyond the immediate area of contamination; however, exposure of PAHs to non-resident organisms should not be discounted. Community-level interactions could occur, through consumption of contaminated benthic organisms by transient fish. The result would be greater exposure to PAHs than estimated assuming the resident benthic and aquatic species are the only receivers of PAH exposure.

Potential for natural recovery

The potential for natural recovery at the Ashland site without intervention is minimal. PAHs are metabolized very slowly in sediments (Su et al. 2000). The most compelling evidence that this is the case at Ashland is the fact that even after many years, PAH concentrations remain high in the sediments. Therefore, ecological effects are not likely to be mitigated by time alone. While some species may have achieved a tolerance to PAHs (some benthic species were apparently abundant in highly contaminated areas) the populations of most benthic organisms at Ashland have been reduced by exposure to the PAHs.

Without remediation, future development of the adjacent shoreline may result in greater exposure of PAHs to organisms. Any process that disturbs sediments, such as dredging or boat traffic, may result in the introduction of sediment bound PAHs from deeper sediments to the water column, where they may be transported to other locations, and undergo photochemical reactions that will result in greater toxicity to aquatic organisms (Bonnet et al. 2000).

References

  1. Albers, P. H. 1995. Petroleum and individual polycyclic aromatic hydrocarbons, p. 330-355. In D. J. Hoffman, B. A. Rattner, G. A. Burton Jr. and J. Cairns Jr. [eds.], Handbook of Ecotoxicology. Lewis Publishers.
  2. Ankley, G. T., R. J. Erikson, G. L. Phipps, V. R. Mattson, P. A. Kosian, B. R. Sheedy, and J. S. Cox. 1995. Effects of light intensity on the phototoxicity of fluoranthene to a benthic microinvertebrate. Environm. Sci. Technol. 29: 2828-2833.
  3. Barron, M. G., E. E. Little, R. Calfee, and S. Diamond. 2000. Quantifying solar spectral irradiance in aquatic habitats for the assessment of photoenhanced toxicity. Environ. Toxicol. Chem. 19: 920-925.
  4. Bestari, K. T. J., R. D. Robinson, K. R. Solomon, T. S. Steele, K. E. Day, and P. K. Sibley. 1998. Distribution and composition of polycyclic aromatic hydrocarbons within experimental microcosms treated with liquid creosote. Environ. Toxicol. Chem. 17: 2359-2368.
  5. Bonnet, C., J. Garric, M. Babut, J.-F. Férard, and L. Martel. 2000. Assessing the potential toxicity of resuspended sediment. Environ. Toxicol. Chem. 19: 1290-1296.
  6. Burkhard, L. P., and M. T. Lukasewycz. 2000. Some bioaccumulation factors and biota-sediment accumulation factors for polycyclic aromatic hydrocarbons in Lake Trout. Environ. Toxicol. Chem. 19: 1427-1429.
  7. Ciarelli, S., B. J. Kater, and N. M. van Straalen. 2000. Influence of bioturbation by the amphipod Corophim volutator on fluoranthene uptake in the marine polychaete Nereis virens. Environ. Toxicol. Chem. 19: 1575-1581.
  8. Clements, W. H., J. T. Oris, and T. E. Wissing. 1994. Accumulation and food chain transfer of fluoranthene and benzo[a]pyrene in Chironomus riparius and Lepomis macrochirus. Arch. Environm. Contam. Toxicol. 26: 261-266.
  9. Couch, J. A., and J. C. Harshbarger. 1985. Effects of carcinogenic agents on aquatic animals: An environmental and experimental overview. Environm. Carcin. Rev. 3: 63-105.
  10. DeLeon, I., J. Ferrario, and C. Byrne. 1988. Bioaccumulation of polynuclear aromatic hydrocarbons by the clam, Rangia cuneata, in the vicinity of a creosote spill. Bull. Environm. Contam. Toxicol. 41: 972-879.
  11. Di Toro, D. M., J. A. McGrath, and D. J. Hansen. 2000. Technical basis for narcotic chemicals and polycyclic aromatic hydrocarbon criteria: I. Water and tissue. Environ. Toxicol. Chem. 19: 1951-1970.
  12. Diamond, S. A., D. R. Mount, L. P. Burkhard, G. T. Ankley, E. A. Makynen, and E. N. Leonard. 2000. Effect of irradiance spectra on the photoinduced toxicity of three polycyclic aromatic hydrocarbons. Environ. Toxicol. Chem. 19: 1389-1396.
  13. Djomo, J. E., P. Garrigues, and J. F. Narbonne. 1996. Uptake and depuration of polycyclic aromatic hydrocarbons from sediment by the zebrafish (Brachydanio rerio). Environ. Toxicol. Chem. 15: 1177-1181.
  14. Elder, J. F., and P. V. Dresler. 1988. Accumulation and bioconcentration of polycyclic aromatic hydrocarbons in a nearshore estuarine environment near a Pensacola (Florida) creosote contamination site. Environm. Poll. 49: 117-132.
  15. Ericson, G., B. Liewenborg, E. Lindesjöö, C. Näf, and L. Balk. 1999. DNA adducts in perch (Perca fluviatilis) from a creosote contaminated site in the River Ångermanälven, Sweden. Aquat. Toxicol. 45: 181-193.
  16. Gensemer, R. W., D. G. Dixon, and B. M. Greenberg. 1998. Amelioration of the photo-induced toxicity of polycyclic aromatic hydrocarbons by a commercial humic acid. Ecotoxicol. Environ. Saf. 39: 57-64.
  17. Gewurtz, S. B., R. Lazar, and G. D. Haffner. 2000. Comparison of polycyclic aromatic hydrocarbon and polychlorinated biphenyl dynamics in benthic invertebrates of Lake Erie, USA. Environ. Toxicol. Chem. 19: 2943-2950.
  18. Ingersoll, C. G., P. S. Haverland, E. L. Brunson, T. J. Canfield, F. J. Dwyer, C. E. Henke, N. E. Kemble, D. R. Mount, and R. G. Fox. 1996. Calculation and evaluation of sediment effect concentrations for the amphipod Hyalella azteca and the midge Chironomus riparius. J. Great Lakes Res. 22: 602-623.
  19. Lampi, M. A., X.-D. Huang, Y. El-Alawi, B. J. McConkey, D. G. Dixon, and B. M. Greenberg. 2001. Occurrence and toxicity of photomodified polycyclic aromatic hydrocarbon mixtures present in contaminated sediments, p.211-221. In B. M. Greenberg, R. N. Hull, M. H. Roberts Jr. and R. W. Gensemer [eds.], Environmental Toxicology and Risk Assessment: Science, Policy and Standardization - Implications for Environmental Decisions: Tenth Volume, ASTM STP 1403. American Society for Testing Materials.
  20. Landrum, P. F., B. J. Eadie, and W. R. Faust. 1991. Toxicokinetics and toxicity of a mixture of sediment-associated polycyclic aromatic hydrocarbons to the amphipod Diporeia sp. Environ. Toxicol. Chem. 10: 35-46.
  21. Lantz, S. E., M. T. Montgomery, W. W. Schultz, P. H. Pritchard, B. J. Spargo, and J. G. Mueller. 1997. Constituents of an organic wood preservative that inhibit the fluoranthene-degrading activity of Sphingomonas paucimobilis strain EPA505. Environm. Sci. Technol. 31: 3573-3580.
  22. Long, E. R., D. D. MacDonald, S. L. Smith, and F. D. Calder. 1995. Incidence of adverse biological effects within ranges of chemical concentrations in marine and estuarine sediments. Environ. Manage. 19: 81-97.
  23. Lowell, R. B., J. M. Culp, and M. G. Dubé. 2000. A weight-of-evidence approach for northern river risk assessment: Integrating the effects of multiple stressors. Environ. Toxicol. Chem. 19: 1182-1190.
  24. Malins, D. C., M. M. Krahn, M. S. Myers, L. D. Rhodes, D. W. Brown, C. A. Krone, B. B. McCain, and S.-L. Chan. 1985. Toxic chemicals in sediments and biota from a creosote-polluted harbour: Relationships with hepatic neoplasms and other hepatic lesions in English Sole (Parophrys vetulus). Carcinogenesis 6: 1463-1469.
  25. Mallakin, A., D. G. Dixon, and B. M. Greenberg. 2000. Pathway of anthracene modification under simulated solar radiation. Chemosphere 40: 1435-1441.
  26. Marwood, C. A., R. E. H. Smith, K. R. Solomon, M. N. Charlton, and B. M. Greenberg. 1999. Intact and photomodified polycyclic aromatic hydrocarbons inhibit photosynthesis in natural assemblages of Lake Erie phytoplankton exposed to solar radiation. Ecotoxicol. Environ. Saf. 44: 322-327.
  27. McConkey, B. J., C. L. Duxbury, D. G. Dixon, and B. M. Greenberg. 1997. Toxicity of a PAH photooxidation product to the bacteria Photobacterium phosphoreum and the duckweed Lemna gibba: effects of phenanthrene and its primary photoproduct, phenanthrenequinone. Environ. Toxicol. Chem. 16: 892-899.
  28. McKee, P., A. Burt, D. McCurvin, D. Hollinger, R. Clement, D. Sutherland, and W. Neaves. 1990. Levels of dioxins, furans and other organic contaminants in harbour sediments near a wood preserving plant using pentachlorophenol and creosote. Chemosphere 20: 1679-1685.
  29. McNally, D. L., J. R. Mihelcic, and D. R. Lueking. 1998. Polycyclic aromatic hydrocarbon degrading microorganisms in Great Lakes sediments. J. Great Lakes Res. 24: 392-403.
  30. Metcalfe, C. D., V. W. Cairns, and J. D. Fitzsimons. 1988. Experimental induction of liver tumors in rainbow trout (Salmo gairdneri) by contaminated sediment from Hamilton Harbour, Ontario. Can. J. Fish. Aquat. Sci. 45: 2161-2167.
  31. Metcalfe, C. D., T. L. Metcalfe, G. Riddle, and G. D. Haffner. 1997. Aromatic hydrocarbons in biota from the Detroit River and western Lake Erie. J. Great Lakes Res. 23: 160-168.
  32. Monson, P. D., G. T. Ankley, and P. A. Kosian. 1995. Phototoxic response of Lumbriculus variegatus to sediments contaminated by polycyclic aromatic hydrocarbons. Environ. Toxicol. Chem. 14: 891-894.
  33. Mueller, J. G., P. J. Chapman, and P. H. Pritchard. 1989. Creosote-contaminated sites: their potential for bioremediation. Environm. Sci. Technol. 23: 1197-1201.
  34. Myers, M. S., L. D. Rhodes, and B. B. McCain. 1987. Pathologic anatomy and patterns of occurrence of hepatic neoplasms, putative preneoplastic lesions and other idiopathic hepatic conditions in English Sole (Parophrys vetulus) from Puget Sound, Washington. J. Nat. Cancer Inst. 78: 333-363.
  35. Oris, J. T., and J. P. Giesy. 1987. The photoinduced toxicity of polycyclic aromatic hydrocarbons to the larvae of the fathead minnow (Pimephales promelas). Chemosphere 16: 1395-1404.
  36. Oris, J. T., A. T. Hall, and J. D. Tylka. 1990. Humic acids reduce the photo-induced toxicity of anthracene to fish and Daphnia. Environ. Toxicol. Chem. 9: 575-583.
  37. Ozretich, R. J., S. P. Ferraro, J. O. Lamberson, and F. A. Cole. 2000. Test of ? polycyclic aromatic hydrocarbon model at a creosote-contaminated site, Elliot Bay, Washington, USA. Environ. Toxicol. Chem. 19: 2378-2389.
  38. Padma, T. V., R. C. Hale, and M. H. Roberts. 1998. Toxicity of water-soluble fractions derived from whole creosote and creosote-contaminated sediments. Environ. Toxicol. Chem. 17: 1606-1610.
  39. Roberts, M., W. Hargis, C. Strobel, and P. F. DeLisle. 1989. Acute toxicity of PAH contaminated sediments to the estuarine fish, Leiostomas xanthurus. Bull. Environm. Contam. Toxicol. 42: 142-149.
  40. Schlueter, M. A., S. I. Guttman, Y. Duan, J. T. Oris, X. Huang, and G. A. Burton. 2000. Effects of acute exposure to fluoranthene-contaminated sediment on the survival and genetic variability of fathead minnows (Pimephales promelas). Environ. Toxicol. Chem. 19: 1011-1018.
  41. Smith, R. E. H., J. A. Furgal, M. N. Charlton, B. M. Greenberg, V. Hiriart, and C. A. Marwood. 1999. Attenuation of ultraviolet radiation in a large lake with low dissolved organic matter concentrations. Can. J. Fish. Aquat. Sci. 56: 1351-1361.
  42. Su, M.-C., E. R. Christensen, J. F. Karls, S. Kosuru, and I. Imamoglu. 2000. Apportionment of polycyclic aromatic hydrocarbon sources in lower Fox River, USA, sediments by a chemical mass balance model. Environ. Toxicol. Chem. 19: 1481-1490.
  43. Suter, G. W. 1996. Toxicological benchmarks for screening contaminants of potential concern for effects on freshwater biota. Environ. Toxicol. Chem. 15: 1232-1241.
  44. Sved, D. W., M. H. Roberts, and P. A. Van Veld. 1997. Toxicity of sediments contaminated with fractions of creosote. Wat. Res. 31: 294-300.
  45. Swartz, R. C., D. W. Schults, R. J. Ozretich, J. O. Lamberson, F. A. Cole, T. H. DeWitt, M. S. Redmond, and S. P. Ferraro. 1995. ?PAH: a model to predict the toxicity of polynuclear aromatic hydrocarbons mixtures in field-collected sediments. Environ. Toxicol. Chem. 14: 1977-1987.
  46. Tagatz, M. E., G. R. Plaia, C. H. Deans, and E. M. Lores. 1983. Toxicity of creosote-contaminated sediment to field- and laboratory-colonized estuarine benthic communities. Environ. Toxicol. Chem. 2: 441-450.
  47. U.S. Environmental Protection Agency. 1998. Guidelines for Ecological Risk Assessment.
  48. van Hattum, B., M. J. C. Pons, and J. F. C. Montanes. 1998. Polycyclic aromatic hydrocarbons in freshwater isopods and field-partitioning between abiotic phases. Arch. Environm. Contam. Toxicol. 35: 257-267.
  49. Vogelbein, W. K., J. W. Fourie, P. A. Van Veld, and R. J. Huggett. 1990. Hepatic neoplasms in the mummichog Fundulus heteroclitus from a creosote-contaminated site. Cancer Res. 50: 5978-5986. Weinstein, J. E., and J. T. Oris. 1999. Humic acids reduce the bioaccumulation and photoinduced toxicity of fluoranthene to fish. Environ. Toxicol. Chem. 18: 2087-2094.
  50. Williamson, C. E., R. S. Stemberger, D. P. Morris, T. M. Frost, and S. G. Paulson. 1996. Ultraviolet radiation in North American lakes: attenuation estimates from DOC measurements and implications for plankton communities. Limnol. Oceanog. 41: 1024-1034.
  51. Zhang, Y., R. S. S. Wu, H.-S. Hong, K.-F. Poon, and M. H. W. Lam. 2000. Field study on desorption rates of polynuclear aromatic hydrocarbons from contaminated marine sediment. Environ. Toxicol. Chem. 19: 2431-2435.

 

The Midwest Hazardous Substance Research Center, Michigan State University.
Please contact us if you have any comments or questions.