Doctor of Philosophy / Water quality guidelines (WQGs) present concentrations of contaminants that are designed to be protective of aquatic ecosystems. In Australia, guidance for assessment of water quality is provided by the ANZECC and ARMCANZ (2000) Guidelines for Fresh and Marine Water Quality. WQGs are generally provided for individual contaminants, not complex mixtures of chemicals, where interaction between contaminants may occur. Complex mixtures of contaminants are however, more commonly found in the environment than singular chemicals. The likelihood and consequences of adverse effects occurring in aquatic ecosystems resulting from contaminants are generally assessed using an ecological risk assessment (ERA) framework. Ecological risk assessment is often a tiered approach, whereby risks identified in early stages, using conservative assumptions, prompt further detailed and more realistic assessment in higher tiers. The objectives of this study were: to assess and investigate the toxicity of the mixture of volatile chlorinated hydrocarbons (VCHs) in groundwater to indigenous marine organisms; to present a ‘best practice’ ecological risk assessment of the discharge of contaminated groundwater to an estuarine embayment and to develop techniques to quantify the environmental risk; and to evaluate the existing ANZECC and ARMCANZ (2000) WQGs for VCHs and to derive new WQGs, where appropriate. Previous investigations at a chemical manufacturing facility in Botany, Sydney, identified several plumes of groundwater contamination with VCHs. Contaminated groundwater containing a complex mixture of VCHs was identified as discharging, via a series of stormwater drains, to surface water in nearby Penrhyn Estuary, an adjacent small intertidal embayment on the northern margin of Botany Bay. A screening level ecological hazard assessment was undertaken using the hazard quotient (HQ) approach, whereby contaminant concentrations measured in the environment were screened against published trigger values (TVs) presented in ANZECC and ARMCANZ (2000). Existing TVs were available for 9 of the 14 VCHs present in surface water in the estuary and new TVs were derived for the remaining 5 VCHs. A greater hazard was identified in the estuary at low tide than high tide or when VCH concentrations from both high and low tides were assessed together. A greater hazard was also identified in the estuary when the toxicity of the mixture was assessed, rather than the toxicity of individual contaminants. The screening level hazard assessment also identified several limitations, including: the low reliability of the TVs for VCHs provided in ANZECC and ARMCANZ (2000); the limited applicability of the TVs to a complex mixture of 14 potentially interacting contaminants; the use of deterministic measures for each of the exposure and toxicity profiles in the HQ method and the associated lack of elements of probability to assess ‘risk’. Subsequent studies were undertaken to address these identified shortcomings of the screening level hazard assessment as described in the following chapters. A toxicity testing methodology was adapted and evaluated for suitability in preventing loss of VCHs from test solutions and also for testing with 6 indigenous marine organisms, including: oyster (Saccostrea commercialis) and sea urchin larvae (Heliocidaris tuberculata); a benthic alga (Nitzschia closterium); an amphipod (Allorchestes compressa); a larval fish (Macquaria novemaculeata); and a polychaete worm (Diopatra dentata). The study evaluated possible VCH loss from 44 mL vials for small organisms (H.tuberculata, S.commercialis and N.closterium) and 1 L jars for larger organisms (M.novemaculeata, A.compressa and D.dentata). Vials were effective in preventing loss of VCHs, however, an average 46% of VCHs were lost from jars, attributable to the headspace provided in the vessels. Test jars were deemed suitable for use with the organisms as test conditions, i.e. dissolved oxygen content and pH, were maintained, however, variability in test organism survival was identified, with some control tests failing to meet all acceptance criteria. Direct toxicity assessment (DTA) of groundwater contaminated with VCHs was undertaken using 5 indigenous marine organisms and site-specific species sensitivity distributions (SSDs) and TVs were derived for the complex mixture of VCHs for application to surface water in Penrhyn Estuary. Test organisms included A.compressa, H.tuberculata, S.commercialis, D.dentata and N.closterium. The SSD was derived using NOEC data in accordance with procedures presented in ANZECC and ARMCANZ (2000) for deriving WQGs. The site-specific SSD adopted was a log-normal distribution, using an acute to chronic ratio (ACR) of 5, with a 95% TV of 838 μg/L total VCHs. A number of additional scenarios were undertaken to evaluate the effect of including different ACRs (i.e. 5 or 10), inclusion of larval development tests as either acute or chronic tests and choice of SSD distribution (i.e. log-normal, Burr Type III and Pareto). TVs for the scenarios modelled varied from 67 μg/L to 954 μg/L total VCHs. A site-specific, quantitative ERA was undertaken of the surface water contaminated with VCHs in Penrhyn Estuary. The risk assessment included probabilistic elements for toxicity (i.e. the site-specific SSD) and exposure (i.e. a cumulative distribution function of monitoring data for VCHs in surface waters in the estuary). The joint probability curve (JPC) methodology was used to derive quantitative estimates of ecological risk (δ) and the type of exposure in the source areas in surface water drains entering the estuary, i.e. Springvale and Floodvale Drains, Springvale and Floodvale Tributaries and the Inner and Outer Estuary. The risk of possible adverse effects and likely adverse effects were each assessed using SSDs derived from NOEC and EC50 data, respectively. Estimates of risk (δ) of possible adverse effects (i.e. based on NOEC data) varied from a maximum of 85% in the Springvale Drain source area to <1% in the outer estuary and estimates of likely adverse effects (i.e. based on EC50 data) varied from 78% to 0%. The ERA represents a ‘best practice’ ecological risk assessment of contamination of an estuary using site-specific probabilistic elements for toxicity and exposure assessments. The VCHs identified in surface water in Penrhyn Estuary are additive in toxicity and act under the narcotic pathway, inhibiting cellular processes through interference with membrane integrity. Lethal toxicity to 50% of organisms (i.e. LC50) is typically reported at the internal lethal concentration (ILC) or critical body residue (CBR) of ~2.5 mmol/kg wet weight or within the range of 1 to 10 mmol/kg wet weight. To evaluate the sensitivity of the test organisms to VCHs and to determine if toxicity in the DTA was due to VCHs, the internal residue for 6 test organisms was calculated for the mixture of VCHs in groundwater and toxicity testing with seawater spiked individually 2 VCHs, chloroform and 1,2-dichloroethane. Calculated residues (at LC50/EC50) were typically between 1 and 10 mmol/kg, with the exception of the algal and sea urchin toxicity tests, which were considerably lower than the expected minimum. Mean internal residues for the groundwater, chloroform and 1,2-dichloroethane were 0.88 mmol/kg, 2.84 mmol/kg and 2.32 mmol/kg, respectively, i.e. close to the predicted value of ~2.5 mmol/kg, indicating that the organisms were suitably sensitive to VCHs. There was no significant difference (P>0.05) between the mean residues of each of the three treatments and the study concluded that the additive toxicity of the VCHs in groundwater was sufficient to account for the observed toxicity (i.e. VCHs caused the toxicity in the DTA undertaken). Evaluation of the existing low reliability ANZECC and ARMCANZ (2000) TVs for chloroform and 1,2-dichloroethane was undertaken to determine if these guidelines were protective of indigenous marine organisms. NOECs, derived from toxicity testing of 1,2- dichloroethane and chloroform with 6 indigenous marine organisms, were screened against the existing low reliability TVs. The TVs for 1,2-dichloroethane and chloroform were protective of 4 of the 6 species tested (A.compressa, D.dentata, S.commercialis and M.novemaculeata), however, the TVs were not protective of the alga (N.closterium) or the sea urchin larvae (H.tuberculata). As the existing TVs were not considered to be adequately protective, SSDs were derived using the NOEC data generated from the testing in accordance with procedures outlined in ANZECC and ARMCANZ (2000). Moderate reliability TVs of 3 μg/L and 165 μg/L were derived for chloroform and 1,2- dichloroethane, respectively, i.e. considerably lower than the existing TVs of 770 μg/L and 1900 μg/L. Differences between the existing and newly derived TVs were considered to result from the sensitive endpoints selected (i.e. growth and larval development rather than survival) and from variability inherent when deriving SSDs using a small number of test species. Ongoing groundwater monitoring indicated that the plumes of VCHs in groundwater, identified in the 1990s, were continuing to migrate towards Botany Bay. Discharge of these groundwater plumes into Botany Bay would result in significant increases in the concentrations of VCHs in the receiving environment and would likely lead to significant environmental impacts. In 2006, a groundwater remediation system was commissioned to prevent the discharge of groundwater containing VCHs into Penrhyn Estuary and Botany Bay. The success of the project had only been measured according to chemical and engineering objectives. Assessment of changes in ecological risk is vital to the success of ERA and central to the ERA management framework. Whereas monitoring of chemical concentrations provides qualitative information that risk should decrease, it cannot quantify the reduction in ecological risk. To assess the ecological risk following implementation of the groundwater treatment system, the risk assessment was revised using surface water monitoring data collected during 2007 and 2008. The ERA indicated that, following remediation of the groundwater, ecological risk in Penrhyn Estuary reduced from a maximum of 35% prior to remediation, to a maximum of only 1.3% after remediation. Using the same methodology applied in the initial risk assessment, the success of the groundwater remediation was measured in terms of ecological risk, rather than engineering or chemical measures of success. Prior to the present investigation, existing techniques for assessing ecological risk of VCH contamination in aquatic ecosystems were inadequate to characterise ecological risk. The current study demonstrated that through monitoring of surface water at the site and DTA using indigenous marine organisms, ecological risk can be assessed using site-specific, quantitative techniques for a complex mixture of VCHs in groundwater. The present investigation also identified that existing ANZECC and ARMCANZ (2000) low reliability TVs were less protective of indigenous test organisms than previously thought and therefore, new TVs were derived in the current work. The present study showed that revision of the risk assessment as conditions change is crucial to the success of the ecological risk management framework.

Doctor of Philosophy / Water quality guidelines (WQGs) present concentrations of contaminants that are designed to be protective of aquatic ecosystems. In Australia, guidance for assessment of water quality is provided by the ANZECC and ARMCANZ (2000) Guidelines for Fresh and Marine Water Quality. WQGs are generally provided for individual contaminants, not complex mixtures of chemicals, where interaction between contaminants may occur. Complex mixtures of contaminants are however, more commonly found in the environment than singular chemicals. The likelihood and consequences of adverse effects occurring in aquatic ecosystems resulting from contaminants are generally assessed using an ecological risk assessment (ERA) framework. Ecological risk assessment is often a tiered approach, whereby risks identified in early stages, using conservative assumptions, prompt further detailed and more realistic assessment in higher tiers. The objectives of this study were: to assess and investigate the toxicity of the mixture of volatile chlorinated hydrocarbons (VCHs) in groundwater to indigenous marine organisms; to present a ‘best practice’ ecological risk assessment of the discharge of contaminated groundwater to an estuarine embayment and to develop techniques to quantify the environmental risk; and to evaluate the existing ANZECC and ARMCANZ (2000) WQGs for VCHs and to derive new WQGs, where appropriate. Previous investigations at a chemical manufacturing facility in Botany, Sydney, identified several plumes of groundwater contamination with VCHs. Contaminated groundwater containing a complex mixture of VCHs was identified as discharging, via a series of stormwater drains, to surface water in nearby Penrhyn Estuary, an adjacent small intertidal embayment on the northern margin of Botany Bay. A screening level ecological hazard assessment was undertaken using the hazard quotient (HQ) approach, whereby contaminant concentrations measured in the environment were screened against published trigger values (TVs) presented in ANZECC and ARMCANZ (2000). Existing TVs were available for 9 of the 14 VCHs present in surface water in the estuary and new TVs were derived for the remaining 5 VCHs. A greater hazard was identified in the estuary at low tide than high tide or when VCH concentrations from both high and low tides were assessed together. A greater hazard was also identified in the estuary when the toxicity of the mixture was assessed, rather than the toxicity of individual contaminants. The screening level hazard assessment also identified several limitations, including: the low reliability of the TVs for VCHs provided in ANZECC and ARMCANZ (2000); the limited applicability of the TVs to a complex mixture of 14 potentially interacting contaminants; the use of deterministic measures for each of the exposure and toxicity profiles in the HQ method and the associated lack of elements of probability to assess ‘risk’. Subsequent studies were undertaken to address these identified shortcomings of the screening level hazard assessment as described in the following chapters. A toxicity testing methodology was adapted and evaluated for suitability in preventing loss of VCHs from test solutions and also for testing with 6 indigenous marine organisms, including: oyster (Saccostrea commercialis) and sea urchin larvae (Heliocidaris tuberculata); a benthic alga (Nitzschia closterium); an amphipod (Allorchestes compressa); a larval fish (Macquaria novemaculeata); and a polychaete worm (Diopatra dentata). The study evaluated possible VCH loss from 44 mL vials for small organisms (H.tuberculata, S.commercialis and N.closterium) and 1 L jars for larger organisms (M.novemaculeata, A.compressa and D.dentata). Vials were effective in preventing loss of VCHs, however, an average 46% of VCHs were lost from jars, attributable to the headspace provided in the vessels. Test jars were deemed suitable for use with the organisms as test conditions, i.e. dissolved oxygen content and pH, were maintained, however, variability in test organism survival was identified, with some control tests failing to meet all acceptance criteria. Direct toxicity assessment (DTA) of groundwater contaminated with VCHs was undertaken using 5 indigenous marine organisms and site-specific species sensitivity distributions (SSDs) and TVs were derived for the complex mixture of VCHs for application to surface water in Penrhyn Estuary. Test organisms included A.compressa, H.tuberculata, S.commercialis, D.dentata and N.closterium. The SSD was derived using NOEC data in accordance with procedures presented in ANZECC and ARMCANZ (2000) for deriving WQGs. The site-specific SSD adopted was a log-normal distribution, using an acute to chronic ratio (ACR) of 5, with a 95% TV of 838 μg/L total VCHs. A number of additional scenarios were undertaken to evaluate the effect of including different ACRs (i.e. 5 or 10), inclusion of larval development tests as either acute or chronic tests and choice of SSD distribution (i.e. log-normal, Burr Type III and Pareto). TVs for the scenarios modelled varied from 67 μg/L to 954 μg/L total VCHs. A site-specific, quantitative ERA was undertaken of the surface water contaminated with VCHs in Penrhyn Estuary. The risk assessment included probabilistic elements for toxicity (i.e. the site-specific SSD) and exposure (i.e. a cumulative distribution function of monitoring data for VCHs in surface waters in the estuary). The joint probability curve (JPC) methodology was used to derive quantitative estimates of ecological risk (δ) and the type of exposure in the source areas in surface water drains entering the estuary, i.e. Springvale and Floodvale Drains, Springvale and Floodvale Tributaries and the Inner and Outer Estuary. The risk of possible adverse effects and likely adverse effects were each assessed using SSDs derived from NOEC and EC50 data, respectively. Estimates of risk (δ) of possible adverse effects (i.e. based on NOEC data) varied from a maximum of 85% in the Springvale Drain source area to <1% in the outer estuary and estimates of likely adverse effects (i.e. based on EC50 data) varied from 78% to 0%. The ERA represents a ‘best practice’ ecological risk assessment of contamination of an estuary using site-specific probabilistic elements for toxicity and exposure assessments. The VCHs identified in surface water in Penrhyn Estuary are additive in toxicity and act under the narcotic pathway, inhibiting cellular processes through interference with membrane integrity. Lethal toxicity to 50% of organisms (i.e. LC50) is typically reported at the internal lethal concentration (ILC) or critical body residue (CBR) of ~2.5 mmol/kg wet weight or within the range of 1 to 10 mmol/kg wet weight. To evaluate the sensitivity of the test organisms to VCHs and to determine if toxicity in the DTA was due to VCHs, the internal residue for 6 test organisms was calculated for the mixture of VCHs in groundwater and toxicity testing with seawater spiked individually 2 VCHs, chloroform and 1,2-dichloroethane. Calculated residues (at LC50/EC50) were typically between 1 and 10 mmol/kg, with the exception of the algal and sea urchin toxicity tests, which were considerably lower than the expected minimum. Mean internal residues for the groundwater, chloroform and 1,2-dichloroethane were 0.88 mmol/kg, 2.84 mmol/kg and 2.32 mmol/kg, respectively, i.e. close to the predicted value of ~2.5 mmol/kg, indicating that the organisms were suitably sensitive to VCHs. There was no significant difference (P>0.05) between the mean residues of each of the three treatments and the study concluded that the additive toxicity of the VCHs in groundwater was sufficient to account for the observed toxicity (i.e. VCHs caused the toxicity in the DTA undertaken). Evaluation of the existing low reliability ANZECC and ARMCANZ (2000) TVs for chloroform and 1,2-dichloroethane was undertaken to determine if these guidelines were protective of indigenous marine organisms. NOECs, derived from toxicity testing of 1,2- dichloroethane and chloroform with 6 indigenous marine organisms, were screened against the existing low reliability TVs. The TVs for 1,2-dichloroethane and chloroform were protective of 4 of the 6 species tested (A.compressa, D.dentata, S.commercialis and M.novemaculeata), however, the TVs were not protective of the alga (N.closterium) or the sea urchin larvae (H.tuberculata). As the existing TVs were not considered to be adequately protective, SSDs were derived using the NOEC data generated from the testing in accordance with procedures outlined in ANZECC and ARMCANZ (2000). Moderate reliability TVs of 3 μg/L and 165 μg/L were derived for chloroform and 1,2- dichloroethane, respectively, i.e. considerably lower than the existing TVs of 770 μg/L and 1900 μg/L. Differences between the existing and newly derived TVs were considered to result from the sensitive endpoints selected (i.e. growth and larval development rather than survival) and from variability inherent when deriving SSDs using a small number of test species. Ongoing groundwater monitoring indicated that the plumes of VCHs in groundwater, identified in the 1990s, were continuing to migrate towards Botany Bay. Discharge of these groundwater plumes into Botany Bay would result in significant increases in the concentrations of VCHs in the receiving environment and would likely lead to significant environmental impacts. In 2006, a groundwater remediation system was commissioned to prevent the discharge of groundwater containing VCHs into Penrhyn Estuary and Botany Bay. The success of the project had only been measured according to chemical and engineering objectives. Assessment of changes in ecological risk is vital to the success of ERA and central to the ERA management framework. Whereas monitoring of chemical concentrations provides qualitative information that risk should decrease, it cannot quantify the reduction in ecological risk. To assess the ecological risk following implementation of the groundwater treatment system, the risk assessment was revised using surface water monitoring data collected during 2007 and 2008. The ERA indicated that, following remediation of the groundwater, ecological risk in Penrhyn Estuary reduced from a maximum of 35% prior to remediation, to a maximum of only 1.3% after remediation. Using the same methodology applied in the initial risk assessment, the success of the groundwater remediation was measured in terms of ecological risk, rather than engineering or chemical measures of success. Prior to the present investigation, existing techniques for assessing ecological risk of VCH contamination in aquatic ecosystems were inadequate to characterise ecological risk. The current study demonstrated that through monitoring of surface water at the site and DTA using indigenous marine organisms, ecological risk can be assessed using site-specific, quantitative techniques for a complex mixture of VCHs in groundwater. The present investigation also identified that existing ANZECC and ARMCANZ (2000) low reliability TVs were less protective of indigenous test organisms than previously thought and therefore, new TVs were derived in the current work. The present study showed that revision of the risk assessment as conditions change is crucial to the success of the ecological risk management framework.