Screening assessment report for chlorinated naphthalenes: chapter 4

Assessment of potential to cause ecological harm

The approach taken in this screening assessment was to examine various pieces of scientific and technical information, and to develop conclusions based on a weight-of-evidence approach and using the precautionary principle as required under section 76.1 of the Canadian Environmental Protection Act, 1999 (CEPA). Since there are unique concerns about di- through octachlorinated naphthalenes (octa-CNs), because they are persistent and bioaccumulative substances, the potential for these homologue groups to cause environmental harm has been evaluated separately from that of monochlorinated naphthalenes (mono-CNs).

Environmental exposure

There is a small dataset of environmental concentrations of CNs in Canada. Much more environmental data on CNs, including data from sediments and soils, have been collected in the U.S. and Europe. CNs have been detected in samples from the following environments in Canada: Arctic and urban air, water from Lake Ontario, fish and birds from the Great Lakes and environs, Pacific coast killer whales, seals and whales from the Canadian Arctic, and Vancouver Island marmots. Environmental exposure data from Canada, the U.S. and elsewhere were tabulated in unpublished working documentation; key data are presented below.

CNs have been measured in air in Canada in Toronto (Harner and Bidleman 1997, Helm and Bidleman 2003), around the Great Lakes and near Cornwall, Ontario (Helm et al. 2003). The highest total-CN concentration measured in Canada was 84.5 pg/m³, measured in suburban Toronto (Helm and Bidleman 2003). The dominant homologue group found over western Lake Ontario and Toronto was tetra-CNs, with smaller proportions of penta- and hexa-CN congeners (Helm et al. 2003; Helm and Bidleman 2003). The percentage of tetra-CNs was significantly higher over Lake Superior than Lake Ontario, while the percentage of penta- and hexa-CNs was higher over Lake Ontario. This is consistent with the greater number of sources of CNs in close proximity to Lake Ontario and the greater influence of transport of the lighter tetra-CN congeners to the Lake Superior region (Helm et al. 2003).

There are multiple studies measuring CNs in the Canadian and other Arctic regions. Harner et al. (1998) and Helm et al. (2004) measured CNs at Alert and Tagish in the Canadian Arctic and found sums of all cogeners chlorinated naphthalene congeners (SCNs) in a concentration of less than 0.01-55 pg/m³; tri- and tetra- CNs were the primary congeners found. Herbert et al. (2005) reports SCN concentrations equal to 27-48 and 9-47 pg/m³ at Ny Ålesund and Tromsø, Norway, respectively.

Lee et al. (2007) measured SCNs, from the Global Atmospheric Passive Sampling survey, in several Arctic regions: Barrow, Alaska, Alert, Nunavut, and Ny Ålesund, Norway. The CN concentrations ranged from 1-8 pg/m³. Tri- and tetra-CNs were the dominant congeners at these Arctic sites. The combustion-related congeners 52/60, 50, 51, 54 and 66/67 were enriched at most sites, indicating the influence of combustion emissions on global levels.

The concentration of CN congeners in Canadian urban air and their transboundary movements was discussed in Harner et al. (2006). The study states that congeners (tri- to octa-CNs) have been found to migrate from Detroit to Point Pelee and Toronto.

Certain CN congeners that are absent in technical Halowax formulations (i.e. CNs 44 and 54) and other congeners that are present in much higher concentrations in combustion sources, such as flue gas and fly ash, than in Halowax mixtures (e.g. CNs 52/60 and 73) are used as indicators of combustion-related CN sources in the environment. Combustion-marker congeners were found to be enriched in the air in suburban and industrial Toronto, as compared to downtown Toronto, suggesting that there are current combustion-related sources in the Toronto suburbs, with evaporative emissions from past Halowax uses being more predominant downtown (Helm and Bidleman 2003).

Although mono- and di-CNs have not been reported in air, this mostly reflects the fact that these homologue groups are typically not measured in air samples.

Few concentrations of CNs have been measured in water, in Canada or elsewhere. In various locations of Lake Ontario, 16.4-24.5 pg/L of tri- and tetra-CNs were measured (Helm et al. 2003). CNs were not detected (at a concentration less than10 ng/L) near an industrial plant in Owen Sound, Ontario, which is on Lake Huron (Kauss 1991), despite the fact that CNs were detected in a concentration of 13 ng/L in the effluent discharged from this facility (Kauss 1991). CNs were also detected sporadically in final effluent discharges on shore at levels of 20-50 ng/L, possibly due to purging of residual contamination from the sewer system. The only other CN measured in water in North America was 2-CN, found in Miami-Dade County, Florida, in concentrations of 740 ng/L (in 2002) and 10 000 ng/L (in 2003), respectively. These measurements may have originated near Superfund sites contaminated with PCBs or near old landfill sites. There are seven Superfund sites in Miami-Dade County (Scorecard 2005).

There is only one known case of an attempt to measure CNs in Canadian sediments. CNs were not detected in surficial sediments (less than 20 ng/g) in the harbour in Owen Sound, Ontario, near an industrial plant (Kauss 1991). The highest identified concentration of CNs in sediment (61 000 ng/g dry wt.) was measured in the Trenton Channel of the Detroit River (Furlong et al. 1988). This level is even higher than that measured near a former chloralkali plant in the state of Georgia, which is known to have caused CN contamination (Kannan et al. 1998). In surficial and suspended sediment samples, penta to hepta-CNs were dominant in most cases.

There are no Canadian soil or sewage sludge data for CNs, and only one measured value in soil was available from the U.S.-that is 17 900 ng/g dry wt., which was measured in soil near a former chloralkali plant in Georgia (Kannan et al. 1998). This concentration is very high compared to values measured in agricultural, residential and industrial soils in England, Spain and Germany (less than 0.01-15.4 ng/g dry wt.) (Meijer et al. 2001, Schumacher et al. 2004, Krauss and Wilcke 2003). However, CNs in a concentration of 1 290 000 ng/g were measured in a municipal waste disposal site in the Netherlands (De Kok et al. 1983).

Only two values were found for CNs in sewage sludge. An average of 83 ng/g dry wt. (range: 50-190 ng/g dry wt.) of di- to penta-CNs were measured in sewage sludge from 14 wastewater treatment plants in the U.K. (Stevens et al. 2003). Concentrations of 3.2 ng/g dry wt. and 5.8 ng/g dry wt. of tetra- and penta-CNs were measured in sewage sludge from a sewage treatment plant in Gothenburg, Sweden (Nylund et al. 1992).

There are few studies of levels of CNs in Canadian biota. No data are available for aquatic invertebrates. Data exist for fish, birds and aquatic mammals from the Great Lakes area and the Arctic. Only one Canadian value exists for terrestrial wildlife (the endangered Vancouver Island marmot). Data are also available for a variety of fish and bird species from Sweden and the Baltic Sea near Poland. The Canadian data, as well as levels in biota in the U.S. and Europe are included in Table 10. Any Canadian data are listed first, with U.S. data second followed by the European data.

In general, tetra-, penta- and hexa-CNs are the dominant homologue groups found in wildlife. The proportions of these congeners found in various species depend on the sources of CNs (i.e. old industrial sources [Halowax] versus combustion sources) and also on the species, since species vary in their ability to metabolize CNs. For example, tetra-CNs dominated (79-82%) in phytoplankton from the Baltic Sea, whereas in herring, the penta-CNs were dominant (50%) (Falandysz and Rappe 1996). In the Canadian Arctic, penta-CNs (33-57%) and hexa-CNs (22-45%) were dominant in beluga whales, whereas in ringed seal, tetra-CNs were dominant (58-83%) (Helm et al. 2002). These differences were attributed to selective metabolism, as is observed in these species with PCDDs (Helm et al. 2002).

Ns: Not stated

Ns: Not stated
^b Only CNs with four to seven chlorines analysed
^e Only CNs with four to six chlorines analysed
^g Only penta- and hexa-CN congeners analysed. No significant time trends evident in the data.

^b Only CNs with four to seven chlorines analysed
^e Only CNs with four to six chlorines analysed

^a Only CNs with four to eight chlorines analysed
^b Only CNs with four to seven chlorines analysed
^d Only CNs with three to seven chlorines analysed
^e Only CNs with four to six chlorines analysed

^e Only CNs with four to six chlorines analysed

Mono-CNs

Risk quotient calculations

Toxicity data for potential receptor organisms were reviewed to find the most sensitive assessment endpoint for each environmentally relevant compartment. A conservative predicted exposure concentration (PEC) was selected for each potential receptor, based on empirical data from monitoring studies. Exposure data for the Canadian environment were used. PECs usually represented reasonable worst-case conditions.

A predicted no-effects concentration (PNEC) was determined, for each assessment endpoint by dividing a critical toxicity value (CTV) by an application factor. CTVs typically represented the lowest acceptable ecotoxicity value. Preference is generally for chronic toxicity data, since long-term exposure is a concern.

Application factors were derived to account for various sources of uncertainty associated with, for example, making extrapolations from acute to chronic effects, from a test species to a different, potentially more sensitive, species, from effects observed in a laboratory to a field setting, and from a single-species test to ecosystems.

Since mono-CNs are relatively soluble in water (see Table 1) and have been detected in water at high concentrations near contaminated sites internationally, risk quotients for pelagic organisms were calculated. Since mono-CNs have not been detected in air or soil internationally, and relevant effects data are lacking, risks to soil biota and terrestrial wildlife through exposure to contaminated air or soil were not considered.

The CTV may be conservatively set at 100 µg/L, which is the lowest concentration of Halowax 1000 (containing from 6.7-69% mono-CN) (see tables 3a and 3b) that significantly reduced the growth of the marine algae Dunaliella teriolecta under chronic exposure conditions (Walsh et al. 1977). The PNEC was determined by dividing the CTV by an application factor of 10 to account for extrapolation from laboratory to field conditions and inter- and intra-species variability. Therefore, the PNEC was 10 µg/L.

Near an industrial plant located in Owen Sound, Ontario, total CNs were detected in final effluent discharge on the shore Owen Sound at levels of 20-50 ng/L (Kauss 1991). However, the CNs identified in the final effluent from the plant were mainly tri- and tetra-CNs closely resembling Halowax 1099 (Kauss 1991). 1-CN and 2-CN were not detected at a method detection limit of 20 ng/L in the final effluent. Nevertheless, for the purposes of this conservative risk quotient, it will be assumed that the final effluent contained 10 ng/L of mono-CNs (one half the detection limit). This effluent discharge will be diluted by a factor of 10 to give a concentration of mono-CNs of 1 ng/L (0.001 µg/L), which will be used as the PEC.

The PEC is divided by the PNEC to estimate the risk quotient. The quotient is therefore calculated as follows:

Quotient for pelagic organisms = PEC / PNEC
= 0.001 µg/L / 10 µg/L
= 0.0001

The quotient is much less than 1, suggesting that there is negligible risk associated with the presence of mono-CNs in Canadian waters.

Other lines of evidence

Mono-CNs are not expected to be persistent in air, or in any other environmental medium. They are furthermore not expected to have significant potential for long-range transport.

Based on available evidence, including in particular the measured log K_ow values and measured BCF values in fish, mono-CNs are not bioaccumulative. However, the available empirical and modelled aquatic toxicity data indicate that mono-CNs may be harmful to aquatic organisms at relatively low concentrations: less than 1 mg/L for acute tests, and 0.1 mg/L for chronic tests.

Mono-CNs have never been detected in the environment in Canada, and CNs have not been in commercial use in Canada for more than two decades.

Di- to octa-CNs

Di- through octa-CNs are persistent in air. Di- through hexa-CNs have been shown to have elevated Arctic contamination potentials. In addition di- through octa-CNs are predicted to be persistent in water and tri- through hepta-CNs> are persistent in both sediment and soil.

Based on the weight of evidence, including in particular measured log K_ow values for di- to octa-CNs, the measured BCF values for di-to penta-CNs in fish, and taking into account the supporting information on measured biomagnification factors (BMFs) for tetra- to hepta-CNs>, the high dietary uptake efficiencies of hexa- to octa-CNs in northern pike, and the very slow elimination of hexa-CNs from the bodies of rats and humans, it is concluded that di- to octa-CNs are also bioaccumulative.

There are special concerns about persistent and bioaccumulative substances that are also harmful to organisms at low concentrations in controlled toxicity tests. Although current science is unable to accurately predict the ecological effects of these substances, they are generally acknowledged to be highly hazardous. Assessments of such substances may thus be performed using a more conservative (precautionary) approach than used for other substances.

Evidence that a substance is very persistent and bioaccumulative, as defined in the Persistence and Bioaccumulation Regulations of CEPA (Canada 2000), when taken together with the potential for environmental release and for toxicity to organisms, provides a significant indication of the substance's potential to cause harmful long-term ecological effects (Environment Canada 2006). Substances that are persistent remain in the environment for a long time, increasing the magnitude and duration of exposure. Releases of small amounts of bioaccumulative substances may lead to high internal concentrations in exposed organisms. Highly bioaccumulative and persistent substances are of special concern, since they may biomagnify in food webs, resulting in very high internal exposures, especially for top predators.

The available empirical and modelled aquatic toxicity data for CNs indicate that di-, tri-, tetra- and penta-CNs may be harmful to aquatic organisms at relatively low concentrations: less than 1 mg/L for acute tests, and 0.1 mg/L for chronic tests. Hexa-, hepta- and octa-CNs were similarly found to cause harmful effects to mammals (particularly cattle) at doses of 2.4 mg/kg body weight day and less.

Beginning around 1910, mono- to octa-CNs were produced commercially as Halowax mixtures for a variety of uses. Although CNs have not been in commercial use in Canada for more than two decades, CNs may be produced unintentionally as a by-product of industrial processes involving heat and/or chlorine, such as waste incineration, cement and magnesium production, refining of metals such as aluminium, and drinking water chlorination.

Finally CNs have been detected in a variety of samples over wide areas of Canada. For example, they have been detected in Arctic and urban air, in water from Lake Ontario, in fish and birds from the Great Lakes and environs, in Pacific coast killer whales, in seals and whales from the Canadian Arctic, and in Vancouver Island marmot.

Based on the lines of evidence presented above, particularly the evidence for persistence, bioaccumulation and potential to cause harm at low exposure values in controlled toxicity tests, and taking into account the limitations of existing quantitative risk estimation methods when applied to such substances-especially data-poor ones such as CNs-and recognizing that, although CNs are no longer in commercial use in Canada, they continue to enter the Canadian environment from unintentional production as well as transboundary movement of air, it is concluded that di- through octa-CNs have the potential to cause environmental harm in Canada.

Sources of uncertainty

The amount of CNs unintentionally emitted into the air during incineration, relative to other sources in Canada is not known. Dyke (1998) notes that while a wide range of persistent organic pollutants can be produced during incineration, the lack of overall industrial combustion monitoring can make it difficult to put the specific operational results into context when trying to establish the relative importance of sources.

The environmental fate of CNs is not well understood, since only a few empirical studies have been performed, on the environmental partitioning and degradation or persistence of CNs, for example. Limited data are available on the solubility of individual CN congeners and Halowax mixtures. Available empirical information was therefore supplemented with model predictions. Since these model predictions were based on representative structures for each homologue group, there may be congeners within each group that show higher or lower rates of degradation.

The data on environmental concentrations of CNs in Canada are limited and incomplete. There are no Canadian soil or sewage sludge data for CNs and very little data exist for the lower chlorinated CNs in other environmental samples in Canada (none exists for mono-CN). The few studies of levels of CNs in Canadian biota do not include aquatic invertebrates.

While data indicate bioconcentration and bioaccumulation of CNs, relatively few studies have been performed to ascertain bioconcentration factors and bioaccumulation factors for the various homologue groups.

For mono-CNs, no chronic aquatic studies were identified for the pure substance, and some of the experimentally derived data do not agree with the modelled data. For di-through hepta-CNs>, no aquatic toxicity studies were identified for any of these congeners alone. Aquatic toxicity of the CN homologue groups was thus modelled using ECOSAR and extrapolated from toxicity testing of Halowax mixtures. This extrapolation of toxicity from a mixture of CN homologues to a single CN homologue group is problematic, since there are uncertainties about the composition of the Halowax mixtures themselves. In addition, mixtures of compounds can have cumulative effects that are additive, antagonistic or synergistic, and hence different from those of the individual compounds. For octa-CNs, effect concentrations were found to be above both the measured and predicted solubilities. No ECOSAR toxicity predictions for octa-CNs were available because the log K_ow value is outside of the acceptable range for this model. Due to the lack of appropriate long-term mammalian studies and the absence of empirical information on effects to soil- and sediment-dwelling organisms, the toxicological behaviour of CNs is experimentally not very well characterized.