Summary of public comments received on the draft screening assessment and risk management scope for cyanides

Comments on the draft screening assessment and risk management scope for cyanides, assessed under the Chemicals Management Plan, were submitted by ArcelorMittal Dofasco, Color Pigments Manufacturers Association, Cyanide Council Regulatory Network Inc, Dominion Colour Corporation, Eldorado Gold Corporation, Embassy of Canada in Turkey, International Cyanide Management Institute, Mining Association of Canada, and Newmont Mining Corporation.

Summarized public comments and responses are provided below, organized by topic:

Methodology

Comment summary 1: There is agreement on the methodology used to derive the predicted no-effect concentration (PNEC), and certain stakeholders are not opposed to the proposed addition to Schedule 1 of the Canadian Environmental Protection Act, 1999 (CEPA).

Response 1: noted.

Comment summary 2: In the screening assessment, total cyanide (CN_T) concentrations (which include three types of cyanide species: strong acid dissociable cyanide (CN_SAD) complexes; weak acid dissociable cyanide (CN_WAD) complexes; and free cyanide (CN_Free)) are compared to a PNEC expressed as CN_Free (hydrogen cyanide (HCN) and cyanide anion (CN^-)). Because of the substantial differences in toxicity and bioavailability of these different cyanide species, direct comparison of CN_T concentrations to the CN_Free PNEC is inappropriate and can lead to erroneous interpretations.

Response 2: The screening assessment acknowledges that the environmental fate and behaviour of cyanides is complex and toxicity of cyanides is mediated by CN_Free. CN_WAD is also considered to represent “biologically available” cyanides (AGDH 2010) since it considers both free cyanide (HCN/CN^-) and weak complexes that may dissociate and release CN_Free at realistic environmental conditions.
Concentrations of CN_Free and CN_WAD are the most direct lines of evidence for comparison to the PNEC. However, the availability of CN_Free and CN_WAD concentrations are limited, or unavailable in the case of the iron and steel sector. Therefore CN_T concentrations were also used as a line of evidence which ranges from potentially conservative to realistic, in cases where photodissociation of strong complexes is more likely.
The screening assessment was updated to provide clarification around the interpretation of comparisons between CN_T and the PNEC.

Comment summary 3: It is conservative to use a PNEC of 1.7 μg CN_Free/L to characterize risk but also more stringent than what is used globally in the regulation of cyanide. Using this low value as the basis for regulation could pose challenges due to the difficulty to routinely measure and have confidence in the accuracy as it relates to environmental concentrations of CN_Free.

Response 3: The PNEC is determined based on the assessment of risk to the environment based on evidence-based analysis performed in the screening assessment.
The Government of Canada takes into consideration various factors when developing instruments and determining appropriate regulatory limits, including the assessment and PNEC value, analytical capability, socio-economic factors, and methods for monitoring substances. The Government of Canada is committed to consulting with affected stakeholders as part of the risk management process
The Metal and Diamond Mining Effluent Regulations (MDMER) is currently the primary tool to manage releases of cyanide from the metal mining sector. The MDMER has prescribed limits for CN_T in effluent at 0.5 mg/L, which came into force on June 1, 2021. The MDMER does not measure CN_Free and does not currently have prescribed limits for CN_T in effluent-receiving waterbodies (i.e., exposure areas). ECCC will continue to monitor to assess the effectiveness of the MDMER and these new limits in managing the risks associated with cyanides. 

Comment summary 4: Where CN_WAD and CN_T are available for the same metal mining site in Figure 7-2 and Table 7-2 in the screening assessment, the CN_WAD concentration is substantially lower (by a factor of 2 to 5) and only marginally above the PNEC. For the aggregate WAD data presented in the screening assessment, in no case did a reference sample below the PNEC end up above the PNEC in the exposed sample.

Response 4: Additional CN_Free and CN_WAD data were extracted for 2016 to 2018, where available. The analysis in the screening assessment has been updated to differentiate CN users and CN non-users in Figures 7-2 and 7-3. There are 6 sites presented in the screening assessment with CN_Free and CN_WAD data included in Table 7-2. With the new data considered, CN_Free concentrations at site 8 provides an example where the reference area mean concentration was below the PNEC while the exposure area mean concentration was above the PNEC. A statement indicating the uncertainties with the CN_Free and CN_WAD data has been added to the screening assessment.

Comment summary 5: The uncertainties associated with comparing CN_T monitoring data to the CN_Free PNEC provide a basis for reconsidering the CEPA section 64(a) conclusion. These chemicals are safely and appropriately used in Canada and existing regulations of cyanide releases are adequate to ensure continued safety of the environment.

Response 5: As required under CEPA, the Government of Canada has applied precaution and weight of evidence while characterizing risk of cyanides. The screening assessment includes conservative approaches when there are uncertainties. Measurements of CN_T were more broadly available than CN_WAD and CN_Free, and were the only concentrations available to characterize ecological risks for the iron and steel sector. Text addressing the interpretation of CN_Free, CN_WAD, and CN_T as respective lines of evidence within a precautionary approach has been added to the screening assessment in section 7.3.

Comment summary 6: The dataset for the metal mining sector used to derive predicted environmental concentrations (PEC) includes a considerable number of measurements below the method detection limit (MDL) (i.e., non-detects), which in many cases were high compared to the PNEC. There is concern that the one-half MDL approach used in the screening assessment may result in an over-estimation of actual concentrations. This approach may have led to an erroneous characterization and quantification of the PEC.
Consultation with a statistician is recommended to provide the most appropriate and scientifically robust approach to analyzing these particular datasets.

Response 6: The use of substitution with one-half the MDL to address measurements below the detection limit, although a standard approach, introduces some uncertainty (may over- or under-estimate PECs). To explore this uncertainty, alternative methods for generating descriptive statistics of PECs for the site-specific analyses (Figure 7-3 of the screening assessment) were applied (Helsel 2012; Bolks et al. 2014). Mean and median PECs were generated for the sites using the alternative method (i.e., Kaplan-Meier, robust regression on order statistics, or maximum likelihood estimation) deemed most appropriate for a dataset given the sample size, percentage of non-detects, and number of MDLs. The mean and median exposure area PECs were mostly equal to or greater than those estimated by the substitution approach. In contrast, the mean and median reference area PECs were more often equal to or less than those estimated by the substitution approach. Consequently, the results generated by the alternative methods tended to show a greater difference in mean and median PECs between exposure areas and corresponding reference areas. Overall, the mean and median PECs generated by the alternative methods are similar to those generated by the substitution method. This suggests that the use of alternative methods here do not reduce the potential over-estimation of PECs. Following these results, the exposure characterization was not updated with the alternative methods.

Comment summary 7: The PNEC of 1.7 μg/L derived in the screening assessment is at the detection limit of the most sensitive analytical methodology approved by the United States Environmental Protection Agency to analyze cyanides in water. This analytical methodology is vulnerable to interferences that may falsely increase or decrease quantities analyzed. Cyanide detection at the PNEC would therefore be vulnerable to the method’s accuracy and precision limitations. The monitoring data presented in the screening assessment does not include a description of the analytical method used to quantitate cyanide. This information is important for determining the reliability of the data, especially since some of the reported concentrations are low.

Response 7: It is acknowledged that the PNEC is on the order of the detection limit of the most common analytical methods for cyanides in water. The screening assessment considers the best data available and applies a conservative approach when there are uncertainties.
Descriptions of analytical methods used to measure cyanides has been added, where available (namely, for the metal mining sector).
Additional discussion regarding the uncertainty associated with elevated detection limits has been added to the screening assessment in section 7.3.3.

Comment summary 8: The sections of the screening assessment that describe persistence and bioaccumulation are inconsistent and contradictory. Additionally, the data provided in the screening assessment does not support the conclusion that CN_Free and its precursors meet the persistence criteria. Consider describing the persistence criteria of the Persistence and Bioaccumulation Regulations of CEPA in the screening assessment. Application of the persistence criteria should be consistent and clearly state and differentiate the species and physical state for which the persistence criteria apply.

Response 8: Information on the persistence (in air, water, and soil) and bioaccumulation of CNFree and precursors to CNFree, including CNWAD and CNSAD, was considered in the screening assessment. A substance that has the characteristics for persistence in at least one compartment is considered persistent (Canada 2000). HCN is considered to be persistent in air.
Text in section 6.2 of the screening assessment has been updated, including additional data on the persistence of iron cyanide species in soil and sediment.

Comment summary 9: The screening assessment states that CN_SAD dissociates in strong acidic conditions. However, strong acidic conditions are not representative of environmental conditions. Many literary sources do not support the claim of rapid dissociation of CN_SAD in the natural environment. Only one reference to photodissociation is provided in the text (Johnson et al. 2002) and this study was performed in relatively warm water which may not be representative of Canadian environmental conditions.

Response 9: Additional references supporting the potential dissociation of ferrocyanides (CNSAD species) were added to the sections for environmental fate and behavior (section 6) and road salts (section 7.2.6) in the screening assessment, namely:

• Broderius and Smith 1980
• Exall et al. 2011 and 2013
• Kuhn and Young 2005
• Young et al. 2005
• EC, HC 2001
• Little and Calfee 2000, 2002, and 2003; and
• Yu et al. 2011

While the upper range of the water temperature in the Johnson et al. (2002) study was indeed relatively warm (14.6°C to 28.2°C), other studies conducted under laboratory conditions at temperatures more representative of natural waters in Canada have observed photodissociation of ferrocyanides (Yu et al. 2011; Broderius and Smith 1980).

Comment summary 10: More information should be included to describe the partitioning of cyanides in water. This includes a discussion of the kinetics of HCN volatilization from liquids at environmentally relevant pH values and calculations to determine how much HCN remains in water versus air.

Response 10: Additional information on the volatilization of HCN from water was added to the screening assessment under environmental fate and behaviour (section 6). Further calculations were not performed, as measured cyanide concentrations used to characterize ecological risks were assumed to represent a relatively steady state between continuous inputs (anthropogenic and natural) and outputs (such as volatilization).

Comment summary 11: It is important to understand the elevated background concentrations of CN_T in the reference areas for metal mines. Factors that result in naturally occurring cyanides in reference areas may also influence concentrations in the exposure areas (e.g., similar hydrological conditions and types of flora and fauna could result in similar natural cyanide concentrations). These elevated background concentrations could be discussed (natural background section).

Response 11: Reference area identification and concentration data were taken directly from the Environmental Effects Monitoring (EEM) dataset. The reference area data for the metal mining sector summarized in Table 7-2, Figure 7-2, and Figure 7-3 of the screening assessment are notably similar to the ambient concentrations summarized in the provincial water quality monitoring datasets when comparing central tendency statistics. Natural sources of cyanides influencing reference area concentrations may also be influencing exposure area concentrations. However, similar CN_T distributions were observed in Figure 7-2 between the exposure area CN_T data for non-cyanide users and those of the reference area CN_T data for cyanide users and non-cyanide users. This observation is discussed in section 7.2.4 where the data appears.

Comment summary 12: The screening assessment would benefit from a summary of data representing natural background concentrations in Canadian waters for both CN_T and CN_WAD and/or CN_Free. It is critical to understand natural background concentrations, in order to ensure the proposed PNEC value is not within the typical background range.

Response 12: Section 7.2.3 of the screening assessment was renamed “ambient concentrations” because the data summarized may not be solely representative of natural background concentrations. Sampling locations in these datasets were not explicitly chosen to collect background concentrations of cyanides but for other monitoring purposes. It is therefore difficult to confidently assess whether the PNEC is within background ranges. Natural background concentrations have not been formally defined for cyanides. As highlighted in other comments, current analytical methodologies are limited in their ability to detect and reliably quantitate cyanides at low concentrations in water. Non-detects were often reported when analyzing ambient samples. The provincial water quality monitoring data summarized in Table 7-1 indicate low percentages of detected concentrations (8% to 45%), as do the reference area data summarized for the metal mining sector (e.g., Figure 7-3, 0% to 50%).

Comment summary 13: The statement suggesting that the occurrence of PNEC exceedances associated with reference area CN_T (and CN_WAD) concentrations “may be due to confounding factors…” should be edited unless further information is available to confirm the presence of former mining sites or tailings sites near the reference areas.

Response 13: Text identifying former mining sites or tailing sites as potential sources of cyanide confounding interpretation of reference area concentrations has been removed from section 7.2.4 of the screening assessment.

Comment summary 14: High CN_T concentrations in the reference areas of metal mines were often assumed to originate from anthropogenic sources, such as road salt application, especially in the northern and western regions of Canada. Such conclusions need to be revisited because road salt is not effective in cold climates and is not typically used. In other instances, such as gold mills in more southern locations, road salts can be a confounding factor.

Response 14: Text was revised to better separate the metal mining and road salt application scenarios. It is agreed that road salt application on highways and roads in proximity to metal mining activities would only be a confounding factor proportional to regional use.

Comment summary 15: The term MDL should probably be replaced with reported detection limit (RDL).
In addition, in the provincial water quality monitoring data for Alberta, 100% of the data have MDLs that exceed the proposed PNEC. This highlights one of the challenges of using CN_T data in the screening assessment, if comparisons to the PNEC will be made.

Response 15: The term MDL has been retained in the screening assessment to maintain consistency and the terminology used by most data providers (i.e., in reporting their achieved MDLs). Additionally, a clear statistical or formal definition of RDL could not be found, and confusion with reporting limits or laboratory reporting limit should be avoided.
Comparisons of provincial water quality data to the PNEC have been removed.

Comment summary 16: It is questioned whether an evaluation of data quality was performed for the monitoring datasets summarized in the screening assessment and whether outlier data were removed following appropriate testing and examination. It is difficult to concur with the statistics provided. In the case of the British Columbia dataset, some MDLs are as high as 100 μg/L. Substitution with one-half MDL for MDLs this high without confirmation of data quality could have an impact on statistical outcomes.

Response 16: The provincial water quality monitoring datasets were assumed to have been quality controlled by the data generators. Datasets were minimally modified in order to be as transparent as possible for reproducible analyses. Suspected outliers and data with high MDLs were not removed but were discussed in text or noted in Table 7-1. However, it was noted that certain sampling sites in the British Columbia dataset were not associated with ambient monitoring (i.e., data labelled “trend”, “permitted discharge”, and “non-permitted discharge”). These sampling locations were removed from the dataset and the summary statistics were updated.

Comment summary 17: The number of facilities with PNEC exceedances identified in Table 7-2 of the screening assessment may not be accurate and should be removed. This is due to the conservative comparison of CN_T PECs to a CN_Free PNEC. The fraction of CN_Free in the CN_T PECs may not exceed the PNEC and therefore the number of facilities with true PNEC exceedances may be lower. Further, it is unclear if the number of facilities represents data points or final discharge points.

Response 17: Table 7-2 was replaced with Figure 7-2 in the screening assessment, and the number of facilities with PNEC exceedances was removed. Figure 7-2 now presents box plots and split violin plots of CN_T concentrations for cyanide users and non-cyanide users by exposure and reference areas. It is acknowledged that the comparison of CN_T concentrations to a CN_Free PNEC is a potentially conservative approach and that the fraction of CN_Free in the CN_T PECs may or may not exceed the PNEC, as discussed elsewhere in this table. Due to limited samples sizes, data from final discharge points were pooled to represent a facility.

Comment summary 18: Descriptive statistics summarizing exposure and reference area CN_T concentrations for the metal mining sector (Table 7-2 in the screening assessment) indicate a large percentage of non-detects (e.g., the exposure and reference area medians are non-detects, as indicated by the footnote). Given this, please add the range of MDLs (as per Table 7-1) and consider rederiving the descriptive statistics using statistical methods more appropriate for datasets with large percentages of non-detects and elevated MDLs.

Response 18: Table 7-2 was replaced with Figure 7-2 in the screening assessment in order to better summarize key information concerning the exposure and reference areas of the metal mining sector, including non-detect information. Figure 7-2 combines box plots illustrating the distribution of CN_T concentrations by exposure and reference areas of cyanides users and non-cyanides users and violin plots showing the distribution and density of detect and non-detect CN_T concentrations.
Alternative statistical methods for generating descriptive statistics were explored for the site-specific analysis of the metal mining sector, but not for the pooled exposure and reference data for the sector. The results were similar between the alternative methods and the substitution method and the means and medians generated by the alternative methods tended to show greater differences in values between exposure and corresponding reference areas. This investigation was not performed for the metal mining sector at large because estimating descriptive statistics using the alternative methods involves estimating values for non-detects based on trends within the original dataset which may not be present when datasets are pooled.
The range of MDLs were added to the accompanying alternative text table for Figure 7-2.

Comment summary 19: There seems to be a bias in the statement on page 26 “Because MDLs (e.g., 0.1 mg/L CN_T) were often higher than the PNEC, a non-detect does not signify the absence of cyanides in a sample and cyanide concentrations exceeding the PNEC may be more numerous than the proportion of detects would suggest”. It is also plausible that cyanide is present at naturally occurring concentrations, or concentrations lower than one-half of MDL, in several cases (as seen in the reference data set).

Response 19:  It is acknowledged that both scenarios are plausible, and the statement has been revised. A non-detect simply indicates the concentration is lower than the detection limit, and the “true” concentration may rest either above or below a substituted one-half method detection limit. Alternative methods to address non-detects were explored, and would reach similar assessment outcomes, as described elsewhere in this table.

Comment summary 20: Provide the detection frequencies for the PEC datasets used in the site-specific analysis for the metal mining sector (i.e., Table 7-3).

Response 20: Percent detection frequencies were added to Figure 7-3 in the screening assessment (formerly Table 7-3). They were also added for other PEC datasets but were unavailable from certain sources which only reported average concentrations.

Comment summary 21: The CN_WAD and CN_Free concentrations in the exposure areas of some sites in Table 7-4 are similar to the corresponding reference concentrations (e.g., Sites 6, 7, and 10). It should be confirmed if this is a function of the data substitution method used, as the number of non-detects is not provided.

Response 21: The detection frequencies were added for sites 1, 2, 6, 9, and A in Table 7-2 of the screening assessment (previously Table 7-4). They were unavailable for sites 1 (2008 to 2010 data) and 5. For sites 1 (2011 to 2016 data), 2, and A, it appears that the similarity in concentrations between the exposure and reference area datasets is due to low detection frequencies. Alternative methods would not be able to estimate central tendencies for most of these datasets due to low detection frequencies (i.e., less than 20%; Helsel 2012) or unknown detection frequencies. Alternative methods were explored for the site 7 datasets to calculate the medians and means for the exposure and reference area (Helsel 2012, Bolks et al. 2014), similar to how they were applied for the datasets in Figure 7-2. The alternative method applied (robust regression on order statistics, rROS) resulted in higher median and mean values than those generated by the substitution approach (one-half MDL). For example, the exposure area median and mean are 11 µg/L and 12 µg/L, respectively, as estimated by rROS, and 2.5 µg/L and 6.7 µg/L, respectively, as estimated by the substitution approach.

Comment summary 22: A table and brief discussion of current detection limits by standard methods of analyses for CN_WAD and CN_Free should be included, ensuring that the PNEC is readily achievable in environmental surface water samples. Specific CN analytical methods are not mandated in Canada. Very few, if any, methods could reliably determine CN_T at 1.7 µg/L in a mining related sample. Bias and interferences are often unquantified when applied to complex matrices such as those in the mining industry. Recent developments through ISO Technical Committee (TC) 147 SC2 resulted in the formation of Working Group 66 “Cyanide methods”, which has standardized two analytical methods that have been able to achieve a 1.5 µg/L limit of quantification for CN_Free in mining related solutions.

Response 22: A review (Ma and Dasgupta 2010) of analytical methods for the quantitation of cyanides indicate the existence of methods with detection limits lower than the PNEC (e.g., 0.5 µg/L, which is considered the lowest reliable detection limit for surface waters; ATSDR 2006; Delaney 2018). Discussion of uncertainties regarding detection limits in the dataset for the metal mining sector was added to section 7.2.4 of the screening assessment, and broader detection limit uncertainties are further discussed in section 7.3.3.

Comment summary 23: The Johnson et al. (2002) paper shows flaws in the analytical chemistry, in addition to the non-representative ambient conditions. The authors also utilize non-standard methodology for the measurement of CN_WAD and CN_T.

Response 23: The text has been updated to acknowledge the uncertainties with this study.

Comment summary 24: Beyond the PEC/PNEC comparison, there is no evidence of observed adverse impacts to aquatic life in Canada documented in the screening assessment.

Response 24: No long-term field studies specific only to cyanide were found in the literature. The screening assessment considers the chronic hazard properties of cyanide and its potential exposures to evaluate ecological risk.

Risk management

Comment summary 25: There should be text in the risk management scope that acknowledges that not all facilities that use cyanide have elevated levels in their effluents and/or receiving environments. There should also be further elaboration on existing management tools, such as provincial permitting requirements, and the EEM provisions of the MDMER, which require monitoring of fish populations.

Response 25: Noted. This is addressed in the risk management approach.

Comment summary 26: Within the risk management scope, sources of literature identify the presence of cyanide in retardants used for fighting forest fires. The risk assessment document does not refer to or evaluate the significance of this potential anthropogenic source of cyanide.

Response 26: The risk management scope does not mention forest fire retardants. However, the screening assessment discusses, in section 4.3, the former use of ferrocyanides in forest fire retardants. Due to the absence of exposure from this source to the environment, this scenario was not considered further.

Comment summary 27: The information gap list requests that CN_WAD data be collected in effluent, which is disagreeable. The PNEC value is a receiving environment value, and effluent data should not be compared to a receiving environment guideline. Therefore, collection of CN_WAD data in effluent for direct comparison with the PNEC value is not appropriate.

Response 27: The information gap list requests that CN_WAD data be collected in effluent, exposure and reference areas in the receiving water bodies. Risk quotients are preferably calculated using receiving environment data; however, effluent data may help provide insight, particularly if receiving environment data are sparse.

Comment summary 28: The number of mines reported to use cyanide should be corrected using data in the “Status Report on the Performance of Metal Mines Subject to the Metal Mining Effluent Regulations in 2015”.

Response 28: Noted. This is addressed in the risk management approach.

Comment summary 29: The proposed wording of the environmental objective should be revised. Presently, it implies that effluent concentrations should meet the PNEC, which is not appropriate or necessary. In addition, effluent concentrations are currently measured as CN_T, whereas the PNEC is based on CN_Free.

Response 29: Noted. This is addressed in the risk management approach.

Comment summary 30: The submitter generally agrees with the proposed risk management objective as written.

Response 30: Noted.

Comment summary 31: Text on existing tools to reduce anthropogenic releases of cyanides to the environment should include current cyanide management practices used in the mining sector.

Response 31: Noted. This is addressed in the risk management approach.

Comment summary 32: The text that describes the use of cyanide in different jurisdictions may cause the reader to incorrectly conclude that use of cyanide is prohibited in various jurisdictions.

Response 32: Noted. This is addressed in the risk management approach.

Comment summary 33: With the reduction in the MDMER effluent limit for CN_T to 0.5 mg/L, and the existing provisions within the MDMER related to both acute lethality and EEM, and the Environmental Code of Practice for Metal Mines, new risk management tools for mining effluent beyond what is in place, are not necessary.

Response 33: The MDMER prescribes a monthly mean limit for CN_T of 0.5 mg/L which came into force on June 1, 2021.   As outlined in the risk management document, no other risk management instrument is proposed at this time to manage releases of cyanide from the metal mining sector. ECCC will continue to monitor to assess the effectiveness of the MDMER and these new limits in managing the risks associated with cyanides.

Comment summary 34: The risk management scope incorrectly states that “approximately 40 percent of measured concentrations of total cyanide in samples collected in areas receiving metal mining effluent exceed the predicted no-effect levels”. The statement should be revised to specify when free and total cyanide data are used, and that reference areas also expressed elevated concentrations above the PNEC.

Response 34: Noted. The statement is revised in the risk management approach.

Comment summary 35: The conclusion that a reduction of the MDMER compliance value for CN_T will reduce CN_Free loading from metal mining sites is not supported by the data presented, and will have significant impact to operations with complex metallurgy.

Response 35: The Metal Mining Effluent Regulations (MMER) were amended in 2018 and renamed the MDMER. Updated deleterious substance effluent limits for new and existing facilities came into effect on June 1, 2021. The maximum authorized monthly mean concentration for CN_T is 0.50 mg/L.

The reduced concentrations of deleterious substances (such as CN_T) in mining effluent are expected to lead to a corresponding reduction in overall loading of deleterious substances such as CN_Free in mine receiving waters. 
The consolidated MDMER is available on the Department of Justice website.

Comment summary 36: New information was provided in response to information requests.

Response 36: Noted.

Comment summary 37: The risk management scope has properly focused on the existing regulatory structure already in place under the MDMER. 

Response 37: Noted.

Comment summary 38: There is concern that the manufacturing and use of colour pigments consisting of multi-metal cyanide compounds (such as copper ferrocyanide) will be negatively impacted by inclusion of all cyanide complexes to Schedule 1 of CEPA, even though these pigments are known to be low risk.

Response 38: Use of colour pigments consisting of multi-metal cyanide compounds was not identified as a source of ecological risk from releases of cyanides in the screening assessment. Adding a substance to Schedule 1 of CEPA does not restrict its use, manufacture, or import. Rather, it enables the Government to take risk management actions under CEPA 1999.

Comment summary 39: Colour pigments, consisting of insoluble multi-iron or multi-metal cyanide containing complexes, should be identified as only soluble in strong acid.

Response 39: The screening assessment describes CN_WAD complexes as including elements such as cadmium, zinc, silver, copper, nickel and mercury that release CN_Free under slightly acidic conditions (pH 4 to 6). It also describes CN_SAD complexes as including elements such as gold, platinum, iron and cobalt that require strong acidic conditions to dissociate. However, CN_SAD complexes may photodegrade to release CN_Free under some environmental settings, including less acidic pH values.

Comment summary 40: Tetrasodium ferrocyanide is completely consumed in an enclosed wet reaction process in the synthesis of copper ferrocyanide. A reasonable exclusion for tetrasodium ferrocyanide should be considered.

Response 40: The screening assessment did not identify the use of tetrasodium ferrocyanide in the synthesis of copper ferrocyanide as a source of cyanides that is of ecological concern. Therefore, n.

Comment summary 41: Contradictory to the risk management scope, Brazil and Turkey have not banned the use of cyanide in gold mining or gold-related production.

Response 41: Noted. This is addressed in the risk management approach.

Comment summary 42: The use of the Cyanide Code as a complimentary risk management tool is a welcome approach. One of the significant challenges is to broaden the Cyanide Code to encompass a greater number of mid-size and smaller mining companies.

Response 42: Noted.

Comment summary 43: The risk management scope concludes that there is cause to lower the national baseline CN_T effluent quality standard contained in the MDMER to reduce the risk to fish habitat. However, there is no comparison provided between MDMER point discharge concentration of CN_T from metal mines, and the aquatic measurements contained in the screening assessment to substantiate this claim.
Metal mining facilities that use cyanides in their process report effluent concentrations under the MDMER and monitor CN_T concentrations in the receiving environment as part of EEM studies. Of the six facilities presented in the screening assessment that have measurements of CN_WAD and CN_Free in areas receiving effluent available, five of the median or average concentrations in the exposure and corresponding reference areas were comparable. However, there are uncertainties for those datasets due to small sample sizes, and, with the exception of one site, low detection frequencies. Monitoring data for CN_T was more readily available by comparison, and CN_T was frequently detected in the receiving environment in concentrations exceeding the PNEC.

Comment summary 44: In the event that an operation has an excess of 1 ppm of CN_T exclusively as iron cyanide, no economically feasible treatment option exists, this has clearly not been taken into consideration.

Response 44: The measures required for mines to achieve compliance with the new effluent limits for deleterious substances such as CN in the MDMER were based on the current performance of the mining industry, existing effluent treatment systems, and submissions made in MDMER EEM reports and/or public data sources (e.g., provincial/ territorial permits).

New information and data

Comment summary 45: A stakeholder indicated using a cyanide-containing substance at low quantity (<0.01 tonnes/year).

Response 45: Noted.

Comment summary 46: The percentage of light at wavelengths below about 400nm is strongly attenuated by the atmosphere, clouds, and in particular plants. For example, very limited diffuse light would reach a pond or stream surface surrounded by tree cover. The remaining light is further attenuated by reflection at the water's surface, and by sorption and scattering within the water column.

Response 46: Additional information on the photodissociation pathway from studies conducted at realistic light and UV intensities was added to the screening assessment.

Comment summary 47: Zepp and Cline (1977) established photo-decomposition rates for chemical species in water, which was applied to iron cyanides by Broderius and Smith (1980) using factors such as water depth, turbidity, season and diurnal fluctuations, who concluded that “HCN formation from the photolysis of iron cyanides is of minimal toxicological importance below a depth of 50-100cm”.

Response 47: The additional information mentioned was considered and the screening assessment was revised.

Comment summary 48: Simovic and Snodgrass (1989) subsequently concluded that CN loss is exclusively controlled through volatilization, as natural waters are usually below pH 8.5; therefore, photo-decomposition of iron cyanide species cannot result in a build-up of CN_Free in aquatic environments.
Mudder (1998) also concludes that decay of iron cyanides in surface waters associated with mining operations, regardless of the surface water configuration or ambient conditions, does not result in the accumulation of CN_WAD.

Response 48: While volatilization is important for the removal of HCN from aquatic environments, the half-life of HCN in water has been observed to vary between 22 hours to 111 hours, depending on various factors including turbulence, initial concentrations and temperatures. Additionally, discussion of two studies (Rader et al. 1995; Osathaphan et al. 2013) has been added to the screening assessment to further discuss the potential longevity of CNFree in solutions exposed to light.      

Comment summary 49: A method is currently being developed for analytical testing of cyanide that has less bias and interference effects than existing distillation-based methods and is directly applicable to mining-related solutions.

Response 49: Noted.

Comment summary 50: Currently underway in Europe is the development of a validated method for the measurement of CN_Free in surface waters. The modifications in this method are directed towards the development of analytical equipment and procedures to measure extremely low concentrations of CN_Free in natural waters.

Response 50: Noted.

Environmental fate and behavior

Comment summary 51: Significant evidence is available to demonstrate that photo-degradation of iron cyanide species proceeds according to first-order kinetics and is the rate limiting step in a multi-species decay path, where volatilization of HCN proceeds at a much faster rate in the environment resulting in an inability for iron cyanide decay to result in accumulation of CN_WAD or CN_Free species in aquatic environments. As presented in the screening assessment, there is no material difference between the reference (background) and the exposed (mine influenced) CN_WAD concentrations demonstrating this rate limiting step for CN_T.

Response 51: The volatilization rate of HCN may be affected by factors such as pH, temperature, water column agitation, surface area to volume of solution, and the initial concentration (Johnson 2015) which may vary by location. Tested in a laboratory, the volatilization half-life of HCN ranges from 22 hours to 111 hours (Broderius and Smith, 1980); therefore, HCN may still accumulate in the environment under certain conditions.
Additional text regarding volatilization rates have been added to the screening assessment.  

Comment summary 52: The screening assessment indicates that CN_SAD will photo-dissociate in the environment to produce CN_Free. However, the only paper that considers an environmental setting is Johnson et al. (2002), measuring photodegradation of Fe(CN)₆ solution from a non-operational heap leach solution discharged in an engineered channel. Johnson et al. (2002) present opposing diurnal fluctuations of CN_WAD and Fe(CN)₆, which they attribute to photo-degradation. This paper presents conclusions at odds with a number of published studies concluding that there are no reported incidents in the mining industry involving discharges of effluents containing elevated levels of iron cyanides, which have resulted in the production of CN_Free levels leading to toxic effects in a receiving system or its aquatic ecosystem (Mudder 1985).

Response 52: The screening assessment was updated to provide further evidence of cyanide complex photodegradation in environmental settings. 
Studies by Exall et al. (2011 and 2013) also demonstrated that ferrocyanides can dissociate to CN_Free once de-icing products are applied to roads and parking lots. Little and Calfee (2000, 2002, 2003) also observed the photodissociation of ferrocyanides (in forest fire retardants) to CN_Free.

Consultation and/or stakeholder engagement

Comment summary 53: There is concern that the moiety-based approach used with the ecological screening assessment will implicate toxicities by virtue of elemental composition.

Response 53: The ecological screening assessment uses a moiety-based approach for risk characterization because various cyanide-containing substances can degrade to produce CN_Free. The proposed risk management measures address the sectors associated with the ecological concerns described in the screening assessment.

Conclusion

Comment summary 54: There is agreement with the screening assessment conclusion that multi-iron containing precursors, including colour pigments with Chemical Abstracts Service Registry Numbers 14038-43-8 and 25869-00-5 do not pose a significant risk in use.

Response 54: Noted. It has been clarified that the risk to human health is considered to be low for the multi-iron cyanide complexes.

References

Bolks A, DeWire A, Harcum JB (Tetra Tech, Inc., Fairfax, VA). 2014. Baseline assessment of left-censored environmental data using R [PDF]. United States: United States Environmental Protection Agency. [accessed 2019 Nov 13].

Broderius SJ, Smith Jr LL (Department of Entomology, Fisheries, and Wildlife, University of Minnesota, St. Paul, MN). 1980. Direct photolysis of hexacyanoferrate complexes: Proposed applications to the aquatic environment [PDF]. Duluth (MN): Environmental Research Laboratory, Office of Research and Development, US Environmental Protection Agency. 61 p. Report No.:EPA-600/3-80-003. Grant No:R805291. [accessed 2020 Aug 13]

Canada. 2000. Canadian Environmental Protection Act, 1999: Persistence and Bioaccumulation Regulations. P.C. 2000-348, 23 March, 2000, SOR/2000-107.

[EC, HC] Environment Canada, Health Canada. 2001. Priority Substances List Assessment Report: Road Salts. Canadian Environmental Protection Act, 1999. Ottawa (ON): Government of Canada. [accessed July 25, 2016].

Exall K, Rochfort Q, Marsalek J. 2011. Measurement of cyanide in urban snowmelt and runoff. Water Qual Res J Can. 46(2):137-147.

Exall K, Rochfort Q, McFadyen R. 2013. Studies of cyanide species in runoff and road salt samples in Ontario, 2010–12: Final report. Burlington (ON): Water Science and Technology Directorate, Environment Canada.

Helsel DR. 2012. Statistics for Censored Environmental Data Using Minitab and R. 2nd ed. Hoboken (NJ): John Wiley & Sons, Inc. 344 p.

Johnson CA, Leinz RW, Grimes DJ, Rye RO. 2002. Photochemical changes in cyanide speciation in drainage from a precious metal ore heap. Environ Sci Technol. 36(5):840-845.

Johnson CA. 2015. The fate of cyanide in leach wastes at gold mines: An environmental perspective. Appl Geochem. 57:194-205.

Kuhn DD, Young TC. 2005. Photolytic degradation of hexacyanoferrate (II) in aqueous media: the determination of the degradation kinetics. Chemosphere. 60(9):1222-1230.

Little EE, Calfee RD (Columbia Environmental Research Center, U.S. Geological Survey, Columbia, MO). 2000. The effects of UVB radiation on the toxicity of fire-fighting chemicals [PDF]. Final report. Missoula (MT): Wildland Fire Chemical Systems, USDA Forest Service. [accessed 2019 May 28].

Little EE, Calfee RD (Columbia Environmental Research Center, U.S. Geological Survey, Columbia, MO). 2002. Environmental persistence and toxicity of fire-retardant chemicals, Fire-Trol^® GTS-R and Phos-Check^® D75-R to fathead minnows [PDF]. Final report No.:ECO-05. Missoula (MT): Missoula Technology and Development Center, USDA Forest Service. [accessed 2019 May 28].

Little EE, Calfee RD (Columbia Environmental Research Center, U.S. Geological Survey, Columbia, MO). 2003. Effects of fire-retardant chemical products on fathead minnows in experimental streams [PDF]. Final report No.: ECO-04. Missoula (MT): Missoula Technology and Development Center, USDA Forest Service. [accessed 2019 May 28].

Ma J, Dasgupta PK. 2010. Recent developments in cyanide detection: a review. Analytica chimica acta. 673(2):117-125

Mudder, T. 1985. Development of site specific discharge criteria through toxicological testing. Proceedings of the Cyanide and the Environment Conference, editing by Dirk van Zyl, Ph.D., Tucson, Arizona. Volume 1. p. 109-117.

Mudder T. 1998. Derivation of aquatic life criteria for total and iron cyanide. The Cyanide monograph, published by Mining Journal Books, London, England, United Kingdom.

Osathaphan K, Ruengruehan K, Yngard RA, Sharma VK. 2013. Photocatalytic Degradation of Ni(II)-Cyano and Co(III)-Cyano Complexes. Water Air Soil Pollut. 224(8):1-7.

Rader WD, Solujic L, Milosavljevic EB, Hendrix J, Nelson JH. 1995. Photocatalytic detoxification of cyanide and metal cyano-species from previous-metal mill effluents. Environ Pollut. 90(3):331-334.

Simovic L, Snodgrass WJ. 1989. Tailings pond design for cyanide control at gold mills using natural degradation. In Gold mining effluent seminar proceedings, Feb 15–16, 1989, Vancouver BC, March 23-24, 1989, Mississauga, ON.

Young TC, Zhao X, and Theis TL. 2005. Chapter 9: Fate and transport of anthropogenic cyanide in surface water. In: Dzombak DA, Ghosh RS, Wong-Chong G, editors. Cyanide in water and soil: chemistry, risk, and management. Boca Raton (FL): Taylor & Francis Group.

Yu XZ, Peng XY, Wang GL. 2011. Photo induced dissociation of ferri and ferro cyanide in hydroponic solution. Environ Sci Tech. 8(4):853-862.

Zepp RG, Cline DM. 1977. Rates of direct photolysis in aquatic environment. Environ Sci Technol. 11(4):359-366.