Health Canada
www.hc-sc.gc.ca
Common menu bar links
Institutional links
Priority Substances List Assessment Report for Releases from Primary and Secondary Copper Smelters and Copper Refineries - Releases from Primary and Secondary Zinc Smelters and Zinc Refineries
3.0 Assessment of "Toxic" under CEPA (Continued)
3.1 CEPA 1999 64(a): Environment
The environmental risk assessment of the PSL substances "Releases from Primary and Secondary Copper Smelters and Copper Refineries" and "Releases from Primary and Secondary Zinc Smelters and Zinc Refineries" is based on the procedures outlined in Environment Canada (1997a).
Because of the evidence of harm from past releases from Canadian copper smelters and refineries and zinc plants (see Sanderson, 1998, for a summary of these effects), an attempt was made to base these assessments of current releases on realistic assumptions and to estimate the probability of adverse effects. The assessments of the impacts of effluents differed in that they were conducted deterministically, and the risk quotients (RQs) derived are in some respects conservative.
As described previously (Section 2.0), an analysis of exposure pathways for individual release constituents and subsequent identification of sensitive receptors were used to select environmental assessment endpoints (e.g., adverse reproductive effects on sensitive fish species in a community). For each combination of endpoint and release constituent, one or more EEVs were determined - expressed either as concentrations of bioavailable chemical species in air, soil or surface water or as rates of deposition of bioavailable forms from air. Whenever possible, EEVs were based on recent empirical data. Models were used to estimate EEVs if suitable empirical data were not available, and as an additional line of evidence to support empirical values.
An ENEV was also determined for each combination of endpoint and release constituent, by dividing a CTV by an application factor. The CTV is typically an estimate of low toxic effects (e.g., EC25) on the most sensitive environmentally relevant species. To increase realism, application factors used in these assessments were very small - usually 1.0 and never more than 2.0. Consequently, when ENEVs are exceeded, there is a significant probability of effects on sensitive organisms. Since all of the constituents of the releases assessed are natural substances - and some are essential micronutrients - care was taken to avoid ENEVs within normal natural concentration ranges. When exposure values were expressed as deposition rates, ENEVs were converted to critical loads. Critical loads are defined as rates of deposition required for contaminants to reach threshold effect values (i.e., ENEVs) in receiving media. Critical loads were estimated probabilistically using appropriate fate and transport models.
Risk was evaluated for each combination of endpoint and release constituent by calculating one or more risk quotients (i.e., EEV/ENEV or EEV/CL). Effects are considered possible if risk quotients for any release constituent exceeded 1.0. In cases where none of the quotients for individual release constituents exceeded 1.0, their combined effects may be considered. Typically, the releases considered contained various mixtures of metals, and their combined effects may be determined by assuming additivity as described below. Because risk quotients for at least one metal exceeded 1.0 at locations close to most of the facilities examined, potential for harmful effects could generally be established without considering additivity.
Joint effect of metal emissions: Of the various models in the literature describing the joint effect of toxicants on organisms, the additivity model is often the most accurate. In cases where it is not the most accurate, it may still be used because, depending upon how it is applied, it may be conservative; that is, it can predict slightly more severe effects than actually occur (Posthuma et al., 1997).
The additivity model is essentially a sum-of-fractions model. The exposure concentration for each toxicant is normalized relative to a standard toxicity endpoint for that toxicant. Typically, this means that for each toxicant, the exposure concentration is divided by a measure of effects, such as an EC25 or an LC50. The resulting risk quotient is sometimes referred to as the toxic unit (TU). Because these are normalized units (i.e., they represent the fraction of exposure required to experience an effect), they can be added, and the sum is an index of possible toxic effect due to exposure to multiple toxicants. In these assessments, since EEVs were divided by ENEVs, the STU becomes the sum across the metals (subscript "i") of the concentration ("C") of each metal in the mixture of contaminants (EEVCi) divided by the ENEV concentration for that metal singly (ENEVCi):
δTU = S(EEV Ci / ENEVCi).
According to this model, when δTU>1, effects are possible.
In cases where effects are expressed as critical loads, δTU based on CL (δTUCL) is a ratio of flux densities to the surface at the receptor location:
δTUCL = δ(EEV Fi / CLFi)
where EEVFi is the flux density ("F") of metal "i" in the mixture of contaminants, and CLFi is the critical load for that metal. In this case, when δTUCL>1.0, effects are possible.
3.1.1 Copper smelters and refineries
3.1.1.1 Releases to air
For the purposes of this assessment, releases from copper smelters and copper refineries are considered together. This was done for two reasons:
- Copper smelters and copper refineries are connected parts of the process of producing metallic Cu. Furthermore, the distinction between smelters and refineries is not always clear. For example, anode casting - the conversion of blister Cu (the impure product of the smelting process) into anodes for use in electrolytic refining - may take place at either a smelter or a refinery. This casting process can result in significant emissions of both metals and SO2.
- Of the three copper refineries included in these assessments, only one is a stand-alone facility. The other two are co-located with a smelter operated by the same company. Thus, environmental receptors are often exposed to releases from copper smelters and refineries together.
3.1.1.1.1 Sulphur dioxide
Ambient SO2: Data for the monitoring of SO2 in the vicinity of copper smelters and refineries were provided by companies and provincial governments. These represent exposure of organisms to ambient SO2 over time scales of the growing season (April to October) as well as over 1 hour. Monitoring data were summarized in Tables 10 and 11, respectively.
The chronic CTV for SO2 is 21 m g/m3 - for slight effects on forest growth for exposure over the growing season. The acute CTV - for injury to sensitive vegetation - is 900 mg/m3 for 1-hour exposure. ENEVs based on these CTVs were estimated to be 10 mg/m3 and 450 mg/m3, respectively. Derivation of these values was discussed in Section 2.4.1.1.1, and the values are summarized in Table 29.
Table 36 summarizes risk information for ambient SO2. As indicated by source attribution information in the first column of this table, of the five facilities (and region, in the case of Sudbury) that include copper smelting/refining, Falconbridge-Kidd Creek is the only one where other metal production operations contribute to SO2 emissions. Sixty-five percent of the SO2 released from the Falconbridge-Kidd Creek facility is attributed to the copper smelter. At the other locations, all SO2 releases are attributed to copper smelting.
In relation to chronic exposure, "growing season average risk quotients" were obtained by dividing the exposure values calculated as averages for the growing season by the ENEV of 10 mg/m3. Values in excess of 1 indicate exposure over the growing season to concentrations greater than those believed to cause no harmful effects. Values greater than 2 indicate exposure to concentrations in excess of those reported to have harmful chronic effects on sensitive vegetation (i.e. the CTV). Quotient values between 1 and 2 may be interpreted to indicate that harmful effects are "possible" for sensitive receptors, while those greater than 2 indicate that effects on sensitive receptors are "likely."
It is clear from Table 36 that risk quotients for chronic exposure of 2 or greater occur frequently within a few kilometres of the facilities, with lower risk quotients observed at greater distances. In general, the entire Sudbury region shows risk quotients in the "possible" effects range (RQ between 1 and 2). At most facilities, all monitoring stations are located very close to source, making estimation of the area impacted difficult.
Table 36 also provides information on acute risk for exposure to ambient SO2 over time scales of 1 hour. The last two columns show the frequencies of exceedence of risk quotients of 1 and 2, respectively, over the growing season. These risk quotients have been calculated by dividing the 1-hour exposure values by the ENEV of 450 mg/m3. There are a moderate number of exceedences of the 1-hour RQ=1 level near all facilities. As may be expected, there are significantly fewer exceedences of the RQ=2 level. The exception to this is one monitoring station located close to the HBM&S facility, where more than half of the 1-hour averages that exceed RQ=1 also exceed RQ=2. The large number of exceedences at monitoring stations close to this facility may indicate that a significant proportion of SO2 releases are from fugitive sources.
It should be noted that most monitoring stations produce about 5000 valid 1-hour averages over the period of the growing season. Therefore, the number of times these "possible" and "likely" acute risk levels are exceeded represent a relatively small percentage of the total time. However, even a single exceedence could cause damage to sensitive plants, as the ENEV is based on a 1-hour exposure.
The "Maximum" column of Table 36 shows the risk quotient for the highest 1-hour average concentration (measured over the growing season) at each monitoring station. These quotients represent the extreme for SO2 exposure for vegetation in the vicinity of copper smelting and refining facilities.
It should be noted that all SO2 monitoring stations near the Noranda-Horne facility are located in the town's residential areas to the northwest, southwest, south and southeast of the smelter. None of the stations is located to the north, northeast or east of the facility, which are the directions typically downwind of the smelter. Risk quotients in these downwind directions would be expected to be somewhat higher.
Quotients for the Noranda CCR facility are not shown in Table 36 because only limited data for SO2 monitoring in the vicinity of the Noranda-CCR facility were available. Further, source attribution is problematic, as this copper refining facility has relatively minor emissions of SO2, while other significant SO2 sources are present in the same geographical area.
- Values in bold meet or exceed a risk quotient of 1.0.
- In some cases, a range is shown for "Growing season average risk quotient," as insufficient data were available to properly correct for values below the detection limit. The lower value is calculated by letting all values below the detection limit equal zero. The higher value is the sum of the lower value and one-half of the detection limit.
Uncertainties: Uncertainties associated with the estimation of exposure to ambient SO2 include the placement of SO2 monitors in locations that may result in overestimation or underestimation of exposure levels typical of the area. An example of underestimation was provided above. Further, the high detection limits of some monitoring instruments necessitated statistical analysis, which likely introduced minor error in estimates of seasonal average SO2 concentrations. There is also uncertainty associated with selection of CTVs and ENEVs, although the effects information base is relative large for ambient SO2.
There is a significant body of evidence of detrimental effects on the environment resulting from fumigations of ambient SO2 in the vicinity of copper smelting facilities. These are mostly related to the high releases of SO2 in the past. In particular, damage to the Sudbury region has been extensively documented (see, for example, Linzon, 1999, and references cited therein).
Based on data in Table 36, it may be concluded that there is the possibility for effects on sensitive vegetation from both acute (1-hour) and chronic (growing season) exposure to SO2 released from the smelting component of copper smelting/refining facilities. Although there are few monitoring stations located further than 3 km from the facilities, data for the Sudbury region indicate that the impacted area may extend to 10 km or more from the source. Distances over which effects on sensitive species are more likely (indicated by risk quotients of greater than 2) are somewhat smaller - generally extending out to 4 km or less from the source, but in some cases extending beyond 10 km.
3.1.1.1.2 Deposited sulphate
Sulphur dioxide emitted from copper smelters and refineries can be oxidized to sulphate in the atmosphere. Both sulphur dioxide and sulphate can be transported long distances from the source, resulting in acidic deposition to soils and lakes over large areas.
The source-receptor model IAM (see Section 2.3.1.1.3) has been used to estimate annual wet sulphate deposition in four regions of eastern Canada. IAM was calibrated to account for oxidation, transport and sulphate deposition based on SO2 emission sources throughout Canada and the United States for the period 1990-1993. The four receptor regions considered are Algoma, Ontario; Muskoka, Ontario; Montmorency, Quebec; and Kejimkujik, Nova Scotia. Estimates of annual total wet sulphate deposition to the four areas from the period 1990-1993 are shown in Table 37.
- Deposition values were produced by IAM based on emission data for the period 1990-1993.
- Critical loads are based on Jeffries et al. (1999).
- Source attributions are relative to the sum of anthropogenic and natural deposition.
Critical loads for wet sulphate deposition derived to allow 95% of lakes to maintain a pH of 6.0 or higher are also shown in Table 37. These are based on evaluation of between 200 and 300 lakes in each of the four regions considered (Jeffries et al., 1999). A critical load for Muskoka was not available. Therefore, the value estimated for Sudbury, located about 150 km northwest of Muskoka, was used.
The risk quotients shown in Table 37 were calculated by dividing the estimated total wet sulphate deposition, due to anthropogenic and natural SO2 sources, by the estimated critical loads for wet sulphate deposition for each of the four areas. At all receptor locations considered, the calculated risk quotient is greater than 1, indicating a potential risk to the environment receiving the deposition. It should be recognized that reductions in SO2 emissions in Canada and the United States have occurred since the 1990-1993 period on which this evaluation is based. Continued study has shown, however, that despite these reductions, critical loads are likely still being exceeded in these regions (Acidifying Emissions Task Group, 1997). Note that these risk quotients are based on all North American sources of SO2.
IAM parameters were scaled based on 1995 emission data for the facilities being considered in these assessments, to estimate incremental contributions to deposition attributable to these sources. Attribution based on source type is shown in Table 37. For example, about 7% of the sulphate deposited at Muskoka is due to SO2 released from Canadian copper smelters. Although these percentages have been calculated based on releases from all sources of wet sulphate deposition between 1990 and 1993, comparison of sulphate deposition for these years to those for 1995 at monitoring sites in eastern Canada suggests that total wet sulphate deposition in these regions changed relatively little between 1990 and 1995 (Table 14). It is furthermore recognized that SO2 emissions from several of the facilities being assessed have been reduced somewhat since 1995, the year on which attribution was based. However, anthropogenic emissions from many sources in Canada and the United States have likely also decreased somewhat. Thus, the relative source attribution percentages presented in Table 37 should be fairly reflective of current conditions.
Uncertainties: There is uncertainty inherent in any modelling exercise, including the detailed evaluation of acid deposition in eastern Canada that led to the source-receptor relationships used in this work (Olson et al., 1983). As pointed out in Section 2.3.1.1.3, however, at sites where comparison was possible, there is good agreement between estimates of acidic deposition and results of monitoring conducted by the OME. There are also uncertainties associated with the estimation of critical loads in the receptor areas, as well as in the comparison of source attribution estimates derived from 1995 emission data with IAM modelling based on the years 1990-1993. As discussed above, however, differences between the two time periods are likely fairly minor.
A significant body of evidence of detrimental effects on the environment resulting from historic acid deposition has been established. In particular, damage to the Sudbury region has been extensively documented (see, for example, Sanderson, 1998, and references cited therein).
There are clearly detrimental effects on lakes in eastern Canada owing to anthropogenic releases of SO2. It may be concluded that Canadian copper smelters contribute a moderate portion of the SO2 leading to this acid deposition (up to 8% at the receptor locations considered). Canadian copper refineries appear to have very minor contributions to acid deposition. It should be recalled, however, that the distinction between emissions from the smelting and refining processes is not always clear. It should also be noted that, based on field studies in the Sudbury region, large emission sources such as copper smelters can contribute a much greater fraction of total sulphate deposited within about 100 km of the source, where dry deposition is a significant factor (Keller and Carbone, 1997).
3.1.1.1.3 Deposited metals
Estimates of annual deposition of the metals Cu, Zn, Ni, Pb, Cd and As, based on monitoring data obtained in the vicinity of copper smelters and refineries, are summarized in Tables 15, 17 and 18. Derivation of critical loads for these metals was discussed in Section 2.4.1.1.3, and annual critical loads were summarized in Tables 31 and 33 for terrestrial and aquatic endpoints respectively.
Table 38 shows risk quotients for metals deposited in the vicinity of copper smelters and refineries. Risk quotients were determined by dividing the exposure (deposition) values by the expected effect (critical load) values. Both deposition and critical load estimates are based on soluble forms of metals. As discussed in Section 2.3.1.2.2, deposition values based on dustfall data are expected to be the most reliable. When data of other types are also available, greater weight is given to dustfall values. Comparison of deposition (and hence risk quotients) for sites located close to facilities where both dustfall and TSP monitoring are conducted indicated that, in general, estimation of total deposition from TSP data underestimated the deposition by a factor of 2-5. The likelihood of TSP-based data to underestimate deposition rates was discussed in Section 2.3.1.2.2.
Table 38 Risk quotients for metal deposition as a function of distance from copper and zinc production facilities
Critical loads used in determination of the risk quotients are the 25th percentile values, for either sandy soils or circumneutral to acidic lakes. If deposition were to be continued at these rates in a typical Shield area until steady state is achieved, 25% of the sandy soils and lakes would be expected to be adversely impacted. Thus, when the risk quotient for a particular monitoring station is equal to 1.0, there is a 25% chance that sandy soils or soft-water (Shield-type) lakes in the vicinity of the station will be adversely affected by the contaminant. Risk quotients above 1.0 indicate that there is a greater chance of observing effects near that station and that effects may be more severe. Critical loads derived for sandy soils typical of those found on the Canadian Shield were used to calculate risk quotients for the Noranda-CCR and Noranda-Gaspé facilities. Although these are not located on the Shield, examination of local surface geology and soils maps (Lajoie, 1954; Fulton, 1996; Service des inventaires forestiers, 1995) indicates that sandy soils occur near each of these facilities, making use of soil critical loads suitable for application at these sites. At all other sites, the more sensitive of soil pore water or surface water critical loads were used.
Emission-based source attribution information is also shown in Table 38. Noranda-Gaspé, Noranda-Horne and Noranda-CCR are stand-alone facilities, and all metal emissions may be attributed to copper smelting or copper refining. Most metal emissions in the Sudbury region are attributable to copper processing, although the Inco-Copper Cliff facility also includes a nickel refinery, which contributes to releases of Cu, Ni, Pb and As. The zinc pressure leaching process used at HBM&S-Flin Flon is reported to have insignificant emissions of metals, and emissions from this facility may be fully attributed to the copper smelter. Between 14% and 92% of metal emissions from the Falconbridge-Kidd Creek facility are attributable to copper processing. It should be noted, however, that because these attributions are based on only a partial inventory of sources (e.g., fugitive releases from tailings areas are not included), the relative contributions of smelters and refineries to total metal deposition rates estimated from monitoring data may be somewhat overestimated.
In general, it may be stated that exceedences of 25th percentile critical loads for Cu extend out to greater distances around copper smelters and refineries than those for other metals. The maximum distance at which an exceedence is observed for Cu is 14.0 km, which, assuming symmetrical deposition patterns, equates to an area of greater than 600 km2. Exceedences for Cu extend out at least 2 km at all facilities examined except Noranda-CCR. Exceedences of 25th percentile critical loads for Zn, Pb, Cd and As are also observed, typically out to distances of 2-4 km from the facilities.
