Report on Biological Toxicity Tests Using Pollution Gradient Studies Sydney Harbour EPS 3/AT/2 - July 2000 Marine Environment Division Environment Canada ENVIRONMENTAL PROTECTION SERIES Sample Number: EPS 3 Categories 1 2 3 4 5 6 7 8 9 Canada, Ottawa, Ontario, K1A 0H3, Canada Regulations/Guidelines/Codes of Practice Problem Assessment and Control Options Research and Technology Literature Reviews Surveys Social, Economic and Environmental Impact Assessments Surveillance Policy Proposals and Statements Manuals New subjects and codes are introduced as they become necessary. A list of EPS reports may be obtained from Environmental Protection Publications, Environmental Protection Service, Environment . Minister of Public Works and Government Services Canada 2001 UP WP Wood Protection/Preservation HA 1 / / Report number with the qualifier EPS 3/HA Subject Area Code Report Category Environmental Protection Series AG AN AP AT CC CE CI FA FP HA IC MA Marine Pollutants MM Mining and Ore Processing NR PF PG PN RA RM Reference Methods SF SP SRM Standard Reference Methods TS TX Subjects Agriculture Anaerobic Technology Airborne Pollutants Aquatic Toxicity Commercial Chemicals Consumers and the Environment Chemical Industries Federal Activities Food Processing Hazardous Wastes Inorganic Chemicals Northern and Rural Regions Paper and Fibres Power Generation Petroleum and Natural Gas Refrigeration and Air Conditioning Surface Finishing Oil and Chemical Spills Transportation Systems Textiles Urban Pollution Report on Biological Toxicity Tests Using Pollution Gradient Studies – Sydney Harbour by B.A. Zajdlik Zajdlik & Associates R.R.3 Rockwood, Ontario N0B 2K0 K.G. Doe Toxicology Laboratory Environment Canada P.O. Box 23005 Moncton, New Brunswick C1A 6S8 L.M. Porebski Marine Environment Division Environment Canada 351 St. Joseph Blvd. Hull, Quebec K1A 0H3 Report EPS 3/AT/2 July 2000 ii National Library of Canada Cataloguing in Publication Data Zajdlik, B.A. (Barry Alan), 1963- Report on biological toxicity tests using pollution gradient studies – Sydney Harbour (Report EPS 3/AT/2) Issued also in French under title : Rapport sur les essais biologiques de toxicité au moyen d’études sur le gradient de pollution – Port de Sydney. ISBN 0-660-18492-3 Cat. No. En49-24/1-50E 1. Environmental toxicology – Nova Scotia – Sydney Harbour. 2. Aquatic biology – Nova Scotia – Sydney Harbour – Methodology. 3. Toxicity testing – Nova Scotia -- Sydney Harbour – Methodology. 4. Waste disposal in the ocean – Environmental aspects – Nova Scotia -- Sydney Harbour. I. Doe, K.G. (Kenneth G.) II. Porebski, L.M. (Linda M.) III. Canada. Environment Canada. IV. Canada. Marine Environment Division. V. Series : Report (Canada. Environment Canada) ; EPS 3/AT/2. 363.739’463’463’097169 C2001-980121-1 QH90.8.T68Z32 2001 Cover photo; Sydney Harbour. Source: P. Topping ©Minister of Public Works and Government Services Canada, 2001 iii Readers’ Comments Comments regarding the content of this report should be addressed to: Linda M. Porebski Marine Sediment Advisor Marine Environment Division Environment Canada 351 St. Joseph Blvd. Hull, Quebec K1A 0H3 Cette publication est aussi disponible en français. Pour l’obtenir, s’adresser à: Publications de la Protection de l’environnement Environnement Canada Ottawa (Ontario) K1A 0H3 Review Notice This report has been reviewed by staff of the Marine Environment Division, Environment Canada, and approved for publication. Mention of trade names or commercial products does not constitute recommendation or endorsement by Environment Canada for use. iv v Abstract Biological toxicity tests (responses included survival, fertilization, growth, photoluminescence, and bioaccumulation), sediment and pore water chemistry, and benthic community structure were examined along a known pollution gradient in Sydney Harbour, Nova Scotia, Canada. Major contaminants were PAHs, PCBs, and heavy metals. Relationships between toxicity tests, chemistry, and benthic structure at the different stations were examined. The primary purpose of the study was to assess Environment Canada's Disposal at Sea Program's interim interpretation criteria for each of the toxicity and bioaccumulation assays. This was done by comparing the pass/fail decisions made using current interpretation criteria, for the assays in the battery with chemical guidelines and with the benthic community structure at the stations. A secondary purpose of the study was to identify additional research needs, or refinements to better use the toxicity test battery within the program. Most of the toxicity tests distinguished between the more contaminated sites and the reference sites, some with very good correlations to the major contaminants or the benthic community, or both, although as a whole the three data sets were not significantly correlated. Non-contaminant factors (ammonia, moisture, grain size, etc.) were often correlated with test responses, although to a lesser degree than with contaminants, suggesting a continued need to measure and assess the relative contribution of these factors to the test results. The choice of reference sites appeared to be critical to whether a station would have "passed or failed" according to program criteria, suggesting that further work on reference site criteria is needed. Amphipod survival, bivalve bioaccumulation, and luminescent bacterial tests, in general, performed well and the interim biological toxicity test interpretation criteria reflected a probable effect level for this data set. The probability of observing toxicity, estimated using mean probable effect level quotients, concurred with the proportion of biological tests actually failing. Echinoid fertilization tests and polychaete growth tests did not appear well correlated with the chemical and benthic results. These tests will require additional research. Other findings of the study were that porewater chemistry may be a valuable addition to the suite of tools used to measure sediment toxicity, and that total PAHs provide a good surrogate for individual PAH measurements. vi Résumé Des essais biologiques de toxicité (portant sur la survie, la fertilisation, la croissance, la photoluminescence et la bioaccumulation), la composition chimique des sédiments et de l’eau interstitielle et la structure de la communauté benthique ont été examinés le long d’un gradient de pollution connu dans le port de Sydney, Nouvelle-Écosse, Canada. Les principaux contaminants étaient les HAP, les BPC et les métaux lourds. Les relations entre les essais de toxicité, la composition chimique et la structure benthique dans les différentes stations ont été examinées. L’étude avait pour but principal d’évaluer les critères provisoires d’interprétation du Programme d’immersion en mer d’Environnement Canada pour chacun des essais de toxicité et de bioaccumulation. Pour ce faire, on a comparé les décisions réussite-échec prises à l’aide des critères actuels d’interprétation, pour la batterie d’essais, les lignes directrices concernant les substances chimiques et la structure de la communauté benthique dans les stations. Son but secondaire consistait à définir les autres besoins en recherche ou les améliorations à apporter pour mieux utiliser la batterie d’essais de toxicité du programme. La plupart des essais de toxicité ont permis de différencier les sites plus contaminés et les sites de référence; pour certains, il existait de très bonnes corrélations avec les principaux contaminants ou les communautés benthiques, ou les deux à la fois, mais en général, les trois ensembles de données n’étaient pas corrélés de façon significative. Les facteurs non contaminants (l’ammoniac, l’humidité, la classe granulométrique, etc. ) étaient souvent corrélés avec les réactions aux essais, quoique moins qu’avec les contaminants, ce qui a semblé indiquer la nécessité continue de mesurer et d’évaluer la contribution relative de ces facteurs aux résultats des essais. Le choix des sites de référence a semblé très important pour déterminer si une station « satisfaisait ou non » aux critères du programme, ce qui a indiqué qu’il fallait peaufiner les critères des sites de référence. En général, les essais portant sur la survie des amphipodes, la bioaccumulation dans les bivalves et les bactéries luminescentes ont donné de bons résultats, et les critères provisoires d’interprétation des essais biologiques de toxicité ont porté à croire qu’il existait une concentration produisant un effet probable pour cet ensemble de données. La probabilité de l’observation d’effets toxiques, calculée à l’aide des quotients moyens des concentrations produisant un effet probable, concordait avec la proportion des essais biologiques se soldant par un échec. Les essais de fertilisation des échinides et de croissance des polychètes n’ont pas semblé bien corrélés avec les résultats pour les analyses chimiques et le benthos. Ces essais nécessiteront une recherche plus poussée. L’étude a aussi démontré que la composition chimique de l’eau interstitielle pouvait être un autre outil valable pour mesurer la toxicité des sédiments et que la mesure des HAP totaux peut remplacer avantageusement celle de chacun de ces composés. vii Table of Contents Abstract....................................................................................................................... v Résumé......................................................................................................................... vi List of Tables ............................................................................................................................... x List of Figures........................................................................................................................ xi Acknowledgements............................................................................................................... xii Abbreviations.................................................................................................................. xiv Introduction...................................................................................................................1 Purpose/Objectives ..........................................................................................................................4 Nomenclature and Conventions.......................................................................................................5 Section 1 1 1.1 1.2 Section 2 Materials and Methods..................................................................................................................6 2 Site Selection ...............................................................................................................................6 2.1 2.2 Sample Collection.....................................................................................................................6 2.2.1 Site Positioning....................................................................................................................6 2.2.2 Sampling Methods ...........................................................................................................................6 2.2.3 Sample Homogenization, Storage, and Transport ...........................................................................7 2.3 Biological Toxicity Test Methods ...................................................................................................7 2.3.1 Data Manipulation ...........................................................................................................................7 2.3.2 East Coast Analyses.......................................................................................................................8 2.3.2.1 Amphipod Toxicity Tests ................................................................................................................8 2.3.2.2 Polychaete Survival/Growth Tests ..................................................................................................8 2.3.2.3 Echinoid Fertilization Inhibition Assay...........................................................................................9 2.3.2.4 Bioaccumulation Test Using Baltic Clam .....................................................................................10 2.3.3 West Coast Analyses .....................................................................................................................10 2.3.3.1 Amphipod Toxicity Tests ..............................................................................................................10 2.3.3.2 Polychaete Survival/Growth Tests ................................................................................................11 2.3.3.3 Marine Photoluminescent Bacterium Acute Toxicity Test ...........................................................12 2.3.3.4 Echinoid Fertilization Inhibition Assay.........................................................................................13 2.4 Benthic Macroinvertebrate Survey................................................................................................14 2.4.1 Data Manipulation .........................................................................................................................14 2.4.2 Data Presentation ...........................................................................................................................14 2.5 Sediment Physical/Chemical Analyses..........................................................................................14 2.5.1 Data Manipulation .........................................................................................................................14 2.5.2 Oxidation Reduction (Redox) Potential, Ammonia, and Sulphide................................................15 2.5.3 Sediment Metal Concentrations.....................................................................................................15 2.5.4 Porewater Metal Concentrations ...................................................................................................15 2.5.5 Sediment Organic Compound Concentrations ..............................................................................15 2.5.6 Porewater Organic Compound Concentrations .............................................................................15 2.5.7 AVS and SEM Metals ...................................................................................................................15 Tests of Hypotheses.....................................................................................................................15 2.6 viii Section 3 Survey Observations and Biological Toxicity Test Results .....................................................16 3 3.1 Biological Toxicity Tests...............................................................................................................16 3.1.1 Acute Tests for Sediment Toxicity Using Marine Amphipods .....................................................16 3.1.2 Sublethal Toxicity Test for Metabolic Inhibition Using a Marine Bacterium ..............................18 3.1.3 Sublethal Test for Sediment Toxicity Using Marine Polychaetes.................................................19 3.1.4 Sublethal Toxicity Test for Echinoid Fertilization Inhibition .......................................................21 3.1.5 Summary Plots and Statistics for Biological Toxicity Tests .........................................................22 3.1.6 Bioaccumulation Test Using Bivalves...........................................................................................23 3.2 3.3 Benthic Macroinvertebrate Survey................................................................................................26 Summary of Available Biological Responses ...............................................................................27 Validity of Test Sediments and Biological Toxicity Tests............................................................29 3.4 3.4.1 Suitability of Test Methods for Marine and Estuarine Amphipods ..............................................29 3.4.2 Suitability of Test Sediments for Comparison with Reference Site..............................................30 3.4.3 Species Specific Criteria for Validity of a Test.............................................................................30 Station Pass/Fail Status..................................................................................................................30 3.5 3.6 Sediment Physical/Chemical Analyses..........................................................................................33 3.6.1 Summary of Available Physical/Chemical Measurements............................................................33 Section 4 4 4.1 Tests of Hypotheses .....................................................................................................................38 H1: Homogeneity of Confounding Factors ...................................................................................38 H2: Predicting Biological Toxicity Test Responses......................................................................43 4.2 4.2.1 Correlation Between Independent Variables .................................................................................43 4.2.2 Automated Model Building for Biological Toxicity Test Responses ...........................................44 4.2.3 Consensus Model Building for Biological Toxicity Test Responses............................................47 4.2.3.1 Amphiporeia virginiana.................................................................................................................47 4.2.3.2 Eohaustorius estuarius ..................................................................................................................47 4.2.3.3 Eohaustorius washingtonianus......................................................................................................48 4.2.3.4 Rhepoxynius abronius....................................................................................................................48 4.2.3.5 Dendraster excentricus..................................................................................................................48 4.2.3.6 Lytechinus pictus ...........................................................................................................................49 4.2.3.7 Boccardia proboscidea..................................................................................................................49 4.2.3.8 Polydora cornuta ...........................................................................................................................49 4.2.3.9 Photoluminescent Bacteria ............................................................................................................49 4.2.3.10 Summary of Consensus Model Building for Biological Toxicity Test Responses ......................50 4.2.4 Model Building for Tissue Bioaccumulation ................................................................................51 H3: Do biological toxicity tests perform consistently within sites?..............................................51 4.3 4.4 H4: Does the suite of biological toxicity tests provide a consistent interpretation of the status of the sediment?......................................................................................................................53 4.4.1 H4a: Do all biological toxicity tests characterize the sample in the same way?...........................53 4.4.1.1 Biological Toxicity Tests Pass/Fail Status Relative to Control Sediments...................................53 4.4.1.2 Biological Toxicity Tests Pass/Fail Status Relative to St. Ann’s Harbour ...................................54 4.4.1.3 Biological Toxicity Tests Pass/Fail Status Relative to Station 12 Reference Site........................55 ix H6: Do the biological toxicity tests indicate an effect when the in-situ benthic 4.4.2 H4b: Do biological toxicity tests rank the stations in the same way?...........................................56 4.5 H5: Do the biological toxicity tests indicate an effect when the TELs or PELs are exceeded? ...57 4.6 4.7 macroinvertebrate community does?.............................................................................................59 H7: Do the three evaluation tools characterize sediments in the same way?................................61 H8: How strongly are the three data sets correlated? ....................................................................61 4.8 4.8.1 Heuristic Comparison of Ordinations............................................................................................64 4.8.2 Comparison of the Three Evaluation Tools...................................................................................64 4.9 H9: Do the interim biological toxicity test interpretation criteria need to be adjusted to match assessments of sediment quality using benthic community structure or SQGs.............................65 Section 5 Summary of Hypothesis Tests ....................................................................................................66 5 Section 6 Discussion ...............................................................................................................................69 6 6.1 Performance Evaluation of Biological Toxicity Tests ..................................................................69 6.1.1 Acute Survival Tests......................................................................................................................69 6.1.2 Sublethal Tests..........................................................................................................................69 6.1.3 Bioaccumulation Tests...................................................................................................................70 6.2 Sediment Physical Chemistry ........................................................................................................70 6.2.1 PELs/TELs and Biological Toxicity Tests ....................................................................................70 6.2.2 Porewater and Sediment Chemistry...............................................................................................70 6.2.3 Total PAHs versus Individual PAH Measurements ......................................................................71 6.3 6.4 Choice of Reference Stations ........................................................................................................71 Reference Stations versus Control Stations...................................................................................73 Comparison of Three Sediment Characterization Methods ..........................................................73 6.5 6.5.1 Interpretation Criteria ....................................................................................................................73 6.5.2 Pass/Fail Decisions ........................................................................................................................74 Further Work ...............................................................................................................................74 6.6 References..................................................................................................................... 76 Appendix A Field Data.......................................................................................................................... 80 Appendix B Raw Data for Toxicity Tests and Bioaccumulation.......................................................... 87 Appendix C Rank Correlations Data................................................................................................... 100 xi List of Figures 1 Sample Locations within Sydney Harbour ..........................................................................................3 2 Percent Survival for Amphiporeia virginiana....................................................................................16 3 Percent Survival for Rhepoxynius abronius.......................................................................................17 4 Percent Survival for Eohaustorius washingtonianus.........................................................................17 5 Percent Survival for Eohaustorius estuarius .....................................................................................18 6 Photoluminescent Bacteria Light Inhibition ......................................................................................18 7 Percent Survival for Polydora cornuta ..............................................................................................19 8 GrowthforPolydora cornuta ............................................................................................................19 9 Percent Survival for Boccardia proboscidea.....................................................................................20 10 Growth for Boccardia proboscidea ...................................................................................................20 11 Percent Fertilization for Dendraster excentricus...............................................................................21 12 Percent Fertilization for Lytechinus pictus ........................................................................................21 13 Scaled Biological Test Responses .....................................................................................................22 14 Percent Survival for Macoma nasuta.................................................................................................23 15 PCB Tissue Concentrations in Macoma nasuta.................................................................................24 16 PAH Tissue Concentrations in Macoma nasuta ................................................................................24 17 Tissue Metal Levels in Macoma nasuta ............................................................................................25 18 Benthic Macroinvertebrate Community Richness .............................................................................26 19 Benthic Macroinvertebrate Community Abundance .........................................................................26 20 Benthic Macroinvertebrate Community Structure Indices ................................................................27 21 Sediment Physical Characteristics .....................................................................................................40 22 Redox Potential Against Stations.......................................................................................................40 23 Porewater pH Against Stations ..........................................................................................................41 24 Porewater Ammonia Against Stations ...............................................................................................41 25 Sediment Ammonia Against Stations ................................................................................................42 26 Sediment Sulphide Against Stations ..................................................................................................42 27 Scores Plot for Benthic Macroinvertebrate Abundances...................................................................60 28 Scores Plot for Sediment Physical/Chemistry Variables ...................................................................63 29 Scores Plot for Biological Test Responses ........................................................................................63 xii Acknowledgements This document was written by Barry Zajdlik of Zajdlik and Associates with assistance from Ken Doe and Linda Porebski of Environment Canada. The project was organized and implemented by Linda Porebski with assistance in the field by Paul Topping and Adrian MacDonald of Environment Canada, Steve Devitt and Eric Garnier from Arenicola Marine, and by John Simic and Floyd Patttengale captain and owner of the “M. V. Katherine G”. Toxicity testing was conducted by Environment Canada’s Ken Doe and Paula Jackman in the Atlantic Region and by Graham van Aggelen and Michelle Fennell in the Pacific and Yukon Region. Benthic community analysis was provided by Pat Pocklington of Arenicola Marine. Advice was given on the collection of ammonia, sulphide, and redox potential by Barry Hargrave, Department of Fisheries and Oceans. Analytical Chemistry was overseen by James Doull and Stewart Yee of Environmment Canada. AVS analysis was overseen by Annette Lawson of Dundas Environmental Consulting Inc. Data analysis and interpretaton was done by Barry Zajdlik. Funding was provided by Jim Osborne of Environment Canada, through the Ocean Disposal Research Fund. Specific sections of the document have also been taken from contributors unpublished reports and are greatfully acknowledged as follows: Steve Devitt. Text in Section 2.2 is extracted from: Devitt, S. 1997. Field acitivity report for collection of marine sediments from Sydney Harbour. Prepared for Ms. L. Porebski, Environment Canada, Marine Environment Division, Ottawa, Ontario. Paula Jackman and Ken Doe, Environment Canada, Atlantic Region. Text in Sections 2.3.2.1 through 2.3.2.4 is extracted from: Jackman, P. and K. Doe. 1997. Toxicity of sediments from Sydney Habour, NS–Results of the 1997 Pollution Gradient Study. (Unpublished Report), Environment Canada, Atlantic Region, Moncton, NB. Michelle Fennell and Graham van Aggelen, Environment Canada, Pacific and Yukon Region. Biological toxicity testing. Text in Sections 2.3.3.1 through 2.3.3.4 is extracted from: Fennell, M. and G. van Aggelen. 1997. Biological assessment of sediments collected from Sydney Harbour, Nova Scotia: Site of the 1997 Pollution Gradient Study conducted by Environment Canada’s National Disposal at Sea Program. (Unpublished Report). Pat Pocklington, Arenicola Marine. Text in Section 2.4.2 is extracted from: Arenicola Marine. 1997. Sample collection and analysis for benthic community structure: Sydney Harbour, NS, July 1997. Many thanks to all the contributors. A special thanks should also go to our reviewers who provided us with many useful comments on both the document and approaches. xiii Scott Carr, U.S. Geological Survey, Biological Resources Division, Corpus Christi, Texas Chris Ingersoll, U.S. Geological Survey, Columbia Environmental Research Centre, Columbia, Missouri Deanna Lee, Environment Canada, Disposal at Sea, North Vancouver, British Columbia Ed Long, National Oceanic and Atmospheric Administration, Coastal Monitoring and Bioeffects Assessment Division, Seattle, Washington Abbreviations AVS Eh EC EPA IC25, IC50 ISQG PAH PEL PCB pH ppt SEM SQG TEL TOC acid volatile sulphides redox potential Environment Canada Environmental Protection Agency (United States) the concentration estimated to cause a 25 or 50% inhibition in response, respectively interim sediment quality guideline polycyclic aromatic hydrocarbon probable effect level polychlorinated biphenyl potenz hydrogen parts per thousand simultaneously extracted metals sediment quality guideline threshold level effect total organic carbon xiv Section 1 Introduction This is the second of two pollution gradient studies by Environment Canada's Disposal at Sea Program, examining the field performance of chemical and biological tools proposed for the assessment of marine sediments destined for disposal at sea. Toxicity and bioaccumulation tests, sediment and pore water chemistry, and benthic community structure were examined along a known PAH pollution gradient in Sydney Harbour, Nova Scotia, Canada. Relationships among these elements at the different stations were examined. The primary purpose of the study was to assess the program’s interim interpretation criteria for the battery of toxicity and bioaccumulation assays. This was done by comparing the assessment of sediment quality using current interpretation criteria, with that obtained using the other assays in the battery, chemical guidelines, and assessment of the benthic community structure at the stations. This report also looks at the functioning of Interim Canadian Sediment Quality Guidelines (CCME, 1999) as chemical benchmarks and at the relevance of other chemical assessment tools, in relation to responses in the toxicity tests and in the benthic community structure. In Canada, disposal at sea is regulated under the Canadian Environmental Protection Act (CEPA, 1999). Before any disposal at sea permit is granted, the material is evaluated according to an international waste assessment framework (EC, 1995c). One of the steps within this assessment process is the characterization of the waste's physical, chemical, and biological properties. Environment Canada is now finalizing a tiered approach for this waste characterization process (EC, 1995c). Tier 1 screening levels could use a set of Canadian Sediment Quality Guidelines at the Threshold Effects Level (TEL), developed from a 1 co-occurrence database in which synoptic chemical and biological information is evaluated in terms of probability of effects. The Threshold Effects Levels denote chemical concentrations at or below which no adverse biological effects are expected. Levels above these criteria would trigger Tier 2 investigations of sediment quality, in the form of toxicity and bioaccumulation testing. The battery selected, includes an acute test with amphipods, three sub-lethal tests and one bioaccumulation test (Table 1) (EC, 1992a, b, c; USEPA, 1993). If the sediments or waste materials pass the toxicity/bioaccumulation tests, open water disposal can be considered. Failure in more than one of the tests (or of the acute test alone) disqualifies the material for open water disposal. Interpretation (pass/fail) criteria for the tests have been proposed for use (Table 9) but require field validation (Stebbing, Dethlefsen, and Carr, 1992). The use of sediment quality values at the Probable Effects Level (PEL) (level above which effects are likely to occur) to reject sediments is possible, but is not being considered at this time. Probable Effects Levels are used in this study, however, as chemical benchmarks to assist with the selection of the sampling stations. The study design followed a sediment triad approach (Chapman, 1992; Stebbing, Dethlefsen and Carr, 1992). The inferences made using each of the triad components may be complementary, contradictory, or uninformative; thus a weight of evidence approach was used to assess the quality of the test sediments in relative terms. It was hoped that this approach could be used to support the toxicity test interpretation criteria being promulgated by Environment Canada. Summary of Biological Tests Being Considered for Regulatory Use Table 1 Test/Species Amphiporeia virginiana Eohaustorius washingtonianus Eohaustorius estuarius Rhepoxynius abronius Microtox® (solid-phase, moisture corrected) Macoma nasuta Dendraster excentricus Lytechinus pictus Boccardia proboscidea Polydora cornuta Factors such as total organic carbon, particle size, depth, ammonia, and sulphide may influence organism responses in lab tests and the in situ benthic community structure. Thus efforts were made to select test stations having similar geophysical and chemical properties, to mitigate the effects of these known confounding factors. The site also needed to include a wide chemical gradient so that stations below the TELs (a reference station), stations between the TELs and PELs (intermediate effects) and above the PELs (effects likely) could be evaluated. Potential Canadian sites in British Columbia, Quebec, and the Atlantic provinces were examined. Two sites were selected for gradient studies, one with a predominantly metals gradient at Belledune Harbour, New Brunswick and one with a predominantly organic contaminant gradient at Sydney Harbour, Nova Scotia. This paper focuses only on the Sydney data. The information on the first study in Belledune Harbour, New Brunswick is available as a technical report (Porebski et al., 1998) or shorter paper (Porebski et al., 1999). 2 Organism Type amphipod amphipod amphipod amphipod bacteria bivalve echinoderm echinoderm polychaete polychaete The major contaminants in Sydney Harbour stem largely from historical coke oven effluent discharge into Muggah Creek at the mouth of the South Arm of the harbour (Matheson et al., 1983). A 1994 site selection study indicated that Sydney Harbour (Figure 1) would provide a suitable PAH gradient based on historical information. Gradients for PCBs, cadmium, zinc, nickel, and copper were also identified. Evidence of PAH bioaccumulation in lobster was found in 1980 and 1981 studies (BEAK Consultants Ltd., 1996). In November 1996, a preliminary study was conducted in the North and South Arms of Sydney Harbour to help select test stations along the gradient. Chemical analysis, particle-size distribution, TOC, ammonia, solid phase Microtox® tests, and Toxichromopad. tests were done on 12 potential test stations and four reference stations. The results showed a clear gradient for PAH (from 196 to <1 µg/g) decreasing with distance from the mouth of Muggah Creek in the South Arm from Stations 1 to 12. The study also showed Microtox® luminescence increasing along the gradient from the most (1) to the least (12) contaminated Test Type sediment sediment sediment sediment sediment sediment porewater porewater sediment sediment Response Percent survival Percent survival Percent survival Percent survival Change in luminescence Percent survival, bioaccumulation Percent fertilization Percent fertilization Percent survival, growth rate Percent survival, growth rate Figure 1 3 Sample Locations within Sydney Harbour Stations. The physical parameters were not homogenous along the gradient. Sediments become coarser grained and lower in TOC along the gradient; however, no better site was available. Depths ranged from 10 to 20 metres but did not follow the gradient. Purpose/Objectives 1.1 This study investigates: • the suitability of the interim interpretation criteria for biological toxicity tests; • the effect of confounding factors (TOC, percent moisture, grain size, redox potential, sediment and porewater ammonia, porewater pH, and sulphide) on biological toxicity test interpretation; • the suitability of recommended species for regulatory use; • the relationship between sediment chemistry and biological toxicity test results; and, • the relationship between the in-situ benthic macroinvertebrate community and biological toxicity test results. A series of hypotheses reflecting the study objectives were generated. The hypotheses address issues arising when interpreting the pass/fail status of sediments using three assessment tools: sediment chemistry, biological toxicity tests, and benthic macroinvertebrate community structure. The hypotheses are: • H1: Are potential confounding factors homogeneous across the stations? This hypothesis is tested to verify that the choice of stations achieved the study design goal of minimizing the effect of known confounding factors such as TOC, particle size, ammonia, and Eh. • H2a: Do confounding factors affect the biological toxicity test response? If the 4 confounding factors vary significantly across stations, then differences in responses may be due to the PAH gradient and/or confounding factors. • H2b: Are the dose responses predictable? Hypotheses 2a and 2b are jointly explored using regression techniques. Those variables most correlated with the biological toxicity test responses are used to develop descriptive models. • H3: Does a biological toxicity test perform consistently at a given site? Significant differences in variability between stations may indicate an inconsistent test, or microscale differences in sediment physical/chemical quality. • H4a: Do all biological toxicity tests characterize the sample in the same way? A concordance between negative biological test responses provides a powerful weight of evidence regarding a potential impact. However, a lack of concordance may indicate that constituents of the battery are providing complementary rather than redundant information which is the raison d’être for a battery. • H4b: Do biological toxicity tests rank the stations in the same way? As the pass/fail criteria dichotomizes the results of toxicity tests, a certain degree of information regarding relative sensitivity is lost. This information may be recovered by analyzing the ranking of the stations. Hypotheses 4a and b address the hypothesis: Does the suite of biological toxicity tests provide a consistent interpretation of the status of the sediment? • H5: Do the biological toxicity tests indicate an effect when the TELs or PELs are exceeded? A lack of agreement between the characterization of sediment using biological toxicity tests and SQGs or ISQGs may necessitate an adjustment of biological toxicity test interpretation criteria. • H6: Do the biological toxicity tests indicate an effect when the in-situ benthic macroinvertebrate community does? A lack of agreement between the characterization of sediment using biological toxicity tests and the in-situ benthic macroinvertebrate community may necessitate an adjustment of biological toxicity test interpretation criteria. • H7: Is there concurrence in the assessment of effect/potential effect when using benthic macroinvertebrate community structure, interim biological toxicity test interpretation criteria, and PELs or TELs? The degree of agreement in the classification of sediment using the three characterization tools is compared to a comparison of the three data sets using the raw data. (see H9) • H8: How strongly are the three data sets correlated? The study design is a gradient design using a sediment quality triad approach. The constituents of the triad are biological toxicity tests, sediment physical/chemistry, and in-situ benthic macroinvertebrate community structure. The degree of correlation between these data sets is explored and compared to the pass/fail characterization of the sediments. • H9: Do the interim biological toxicity test interpretation criteria need to be adjusted to match assessments of sediment quality using benthic community structure or PELs or TELs? Should the previous hypothesis tests indicate that the interim biological toxicity test interpretation criteria do not characterize sediments in the same way as in-situ benthic community structure and PELs or TELs; an adjustment to the interim interpretation criteria may be explored. 5 1.2 Nomenclature and Conventions At times, the phrase “along the gradient” is used. This phrase refers to the following ordering of stations: Control, 1, 5, 6, 9, 12, and St. Ann’s Harbour. The “top of the gradient” refers to the inner most stations in Sydney Harbour. Laboratory replicates are synonymous with subsamples and field replicates are synonymous with true replicates or simply replicates. Laboratory replicates are not used in any analyses unless explicitly stated. Wherever possible, field replicates are used. This is not possible when varying levels of replication occur in the same data set. This is the case with combined physical/chemical data sets. Porewater variables were usually only measured once, while and sediment variables were usually measured more often. The use of field replicates is also precluded when sediment variables cannot be identically matched with the samples used to conduct biological toxicity tests. The test’s degrees of freedom are included with all tests in order to clarify the level of replication used in the analysis. Sydney Harbour sediments in refrigerated storage. Source: K. Doe Section 2 Materials and Methods Site Selection 2.1 This study and the Belledune Harbour study (Porebski et al., 1998; 1999) were conducted to evaluate tools used in assessing the suitability of sediments for ocean disposal. The two areas (Belledune and Sydney Harbour) were chosen for their known metals and organic compound gradients, respectively (BEAK Consultants Ltd., 1996). The criteria for site selection within the harbours included homogeneity of noncontaminant factors (i.e., depth, temperature, sediment grain sizes, salinity, TOC, ammonia) and suitability of sediments for testing with the biological tests being evaluated. The reference stations were chosen with the same criteria in mind, and also for the absence of the contaminants being investigated. In the present study, samples were collected from five sites within Sydney Harbour and from a single site within St. Ann’s Harbour (see Figure 1). Sample Collection 2.2 The text in Section 4.2 was extracted from the field report by Devitt (1997). The field trip to collect sediment samples and benthic macroinvertebrates was conducted between July 10th and 12th , 1997. 2.2.1 Site Positioning. Sites were located using a using a Garman 75™ handheld global positioning system (GPS) and marked with a small, anchored buoy. After anchoring, site coordinates were again recorded with the Garman 75™ GPS as well as a Trimble™ handheld GPS for verification. The site coordinates are tabulated in Appendix A, Table A-1 and depicted in Figure 1. 2.2.2 Sampling Methods. Water quality variables including depth, temperature, 6 conductivity, salinity (calculated), pH, dissolved oxygen, and redox potential were measured at 3-min intervals to a depth of 12 m and at 0.5 m above the sediments (see Appendix A; Table A-2). A Hydrolab Water Quality Monitoring System was calibrated according to the manufacturers recommendations using a certified calibration standard for conductivity, certified grade pH buffers for pH, air and a manufacturer-supplied calibration table for dissolved oxygen, and a thermometer calibrated against a National Standards Board thermometer for temperature. A 0.25 m2 Van-Veen type grab sampler was used to collect sediments. After being winched to the surface, the overlying water was poured off and sediment placed in a plastic fish tote. Visual observations of consistency, benthic macroinvertebrate organisms, and odour were recorded. Temperature, redox potential, and pH were measured in the upper 5 cm of sediment. An alcohol-filled field thermometer was used to measure the temperature after stabilization. Redox potential and pH were measured using a Barnant 20™ digital pH/ORP/mV meter, calibrated according to the manufacturer’s recommendations, using certified grade pH buffers. Van-Veen grabs used to collect sediment for toxicity and chemical testing. Source: P. Topping Sampling in Sydney Harbour. Source: P. Topping To ensure the greatest possible homogeneity, grabs for benthos, toxicity tests, and physicochemistry were taken alternately. The benthic macroinvertebrate community was sampled using a 0.1 m2 Van-Veen grab sampler. Five grabs were taken at each of the sites. The sediment and overlying water was placed in a fish tote and then sieved through a 0.5-mm nylon screen. The screened material and benthic macroinvertebrates were preserved in buffered formaldehyde. 2.2.3 Sample Homogenization, Storage, and Transport. Collected toxicity samples were manually pre-mixed using a stainless steel spoon. Samples were aggregated until approximately 60 L of sediment were collected for each of three replicates at each site. The replicate sample was homogenized using a ¾-inch, two-speed drill with a stainless steel paddle. The mixed samples were transferred to containers, pre-labelled with blind sample numbers and were stored in either coolers with gel packs or a large Xactit box with ice for the balance of the working day. Blind sample numbers were as in Appendix A; Table A-3: Blind Sample Numbers. Samples of control sediment (sediment where test organisms were collected or reared) were also taken for analysis by each lab. After each working day, the samples were stored in a walk-in refrigerator at a temperature of 7 2–3oC, at Highland Fisheries Ltd., in Glace Bay, Nova Scotia. Samples destined for chemical and toxicity analyses were shipped from Sydney to Moncton (Mr. K. Doe, Environment Canada Laboratory, Moncton) and to North Vancouver (Mr. S. Yee, Environment Canada Laboratory, North Vancouver). Samples were shipped using refrigerated transport. Samples sent to Moncton were received within three days of shipping but those shipped to North Vancouver were delayed and arrived in a frozen condition 13 days after being shipped. In order to enable testing of fresh sediments in both east and west coast laboratories, the Moncton laboratory sub-sampled each sample they received and shipped those sub-samples by air cargo to the laboratory in North Vancouver to replace the frozen sediments. All tests and analyses were performed using sediments that had never been frozen. Samples collected for AVS analyses were shipped by air cargo to Burlington (Ms. Annette Lawson, Dundas Environmental Consulting, Inc., Burlington, Ontario). Biological Toxicity Test Methods 2.3 2.3.1 Data Manipulation. Biological toxicity test data were manipulated as follows: • non-detected tissue PCB values are replaced with the detection limit of 0.48 ng/g PCBs); • tissue PAH values below the detection limit are replaced by the sample-specific detection limit; • the tissue metal means for Stations 1, 6 and 12 are estimated using two laboratory replicates while tissue metal means for Stations 5, 9, and St. Ann’s Harbour are estimated using three laboratory replicates; • a note on the original data spreadsheets indicates there may be an error in Station 6 tissue metal levels, but does not state what it might be; the data is used as is; and, were 28–30 days old at the start of the test on July 30, 1997. Samples numbered: 6, 36, 57, 66, 68, 92 as well as the control sample number 32 were analyzed. Juvenile B. proboscidea of 30–32 days old were used for the test performed on August 15, 1997. Juveniles were obtained from laboratory cultures originally supplied by Environment Canada, Vancouver, BC and maintained in the laboratory for several years. Samples numbered 35, 66, 68, 92, as well as the control sample number 4 were analyzed. On the day before test initiation, each 4-L bucket of test sediment was homogenized and 175-mL portions were added to each of five, 1-L glass mason jars. The jars were then filled with 800-mL of clean seawater (salinity 28 ± 2 ‰), covered, then aerated overnight with oilfree compressed air at a rate of approximately 150 mL/min. Tests were conducted according to the draft protocol (EC, 1995a). The following day, polychaetes were removed from their holding sediment. For P. cornuta, five animals were added to each of the test vessels. For B. proboscidea, only four animals were used due to an inadequate supply of juveniles. Several juveniles were taken at the start of the tests and washed and dried at 60°C to determine the initial weight of animals. Photoperiod for the testing was 16 hours of light and 8 hours of dark; and salinity was approximately 30 ± 2 ‰. Temperature was maintained at 23 ± 1° C throughout the testing. Animals were fed three times a week with a 1:1 mixture of finely ground Enteromorpha (green marine macro-alga) and Tetramin® (commercial fish flakes) at a rate of 5 mg per worm. Tests were monitored daily for temperature, aeration, and observations. Three times a week, a replicate of each sample was checked for pH, dissolved oxygen, temperature, and salinity. Approximately 80% of the water overlying the test sediment was renewed on day seven. The tests were terminated at 14 days and the contents of each jar were sieved through a 0.5- mm sieve. Any polychaetes not found at test termination were considered dead. Immobile 9 animals were checked under a dissecting scope to confirm death. All surviving polychaetes were washed, dried, and weighed. The mean percentage survival of polychaetes in all the replicates was calculated. Mean weights for the five replicates of each treatment were compared to the mean weights of the control worms using the Sigma Stat Statistical Program (Version 1, Windows, 1994) from Jandel Scientific Software. A reference toxicant test was conducted with CdCl on the P. cornuta using water only exposures for 96-h. Using the survival data at each test concentration, the 96-h LC50 was calculated using the methods of Stephan (1977). Due to an inadequate supply of B. proboscidea, no reference toxicant test was performed. 2.3.2.3 Echinoid Fertilization Inhibition Assay. White sea urchins, L. pictus, tested during the study were from the EP Laboratory stock (received from Marinus Inc. of Long Beach, California, USA, in 1994 and 1996). Testing was performed on July 23, 1997. Samples numbered 2, 6, 11, 15, 23, 24, 33, 34, 38, 40, 48, 71, 73, 75, 80, 84, 87, and 100 were analyzed. Two, 250-mL portions of each sediment were centrifuged at 3000 rpm for 15 min. The supernatant liquids from the replicates were combined and centrifuged for an additional 15 min at 3000 rpm. Porewater was measured for temperature, pH, salinity, and dissolved oxygen. Four replicates of a dilution series for each porewater were prepared. The test was conducted according to Environment Canada (1992b). Sea urchins were injected with 1 mL of 0.5 M potassium chloride (KCl) solution to induce spawning. Eggs produced from all females were pooled, and the concentration was adjusted to 2000 eggs/mL. Sperm were pooled from all males using the "dry" spawning technique, then stored in a vial on ice. A fixed "sperm-to-egg ratio" of 20 000:1 was used to produce approximately 90% fertilization in the controls. Sperm were activated immediately before test initiation. Test volume was 10 mL and test temperature was 20 ± 1oC. Sperm were exposed to the test solutions for 10 min, followed by an additional 10 min-exposure of the sperm and eggs. The test was then terminated using 2 mL of 10% formalin per replicate. One hundred eggs were examined from each replicate to determine the percentage fertilized. Percent fertilization data in each concentration of porewater were used to calculate the IC50 and IC25 for each sediment porewater tested. The linear interpolation method as implemented in the ICPIN program of Norberg-King (1993) was used to estimate the endpoints. A reference toxicant test was performed simultaneously with the porewater toxicity tests using copper sulphate (CuSO4). IC50s were calculated using the linear interpolation method or ICPIN (Norberg-King, 1993). 2.3.2.4 Bioaccumulation Test Using Baltic Clam. M. nasuta were purchased from A.K. Siewers of Santa Cruz, California, USA. The animals were received at EQL, Moncton, NB on July 17, 1997 at 22.5° C. The animals were placed in trays containing aerated seawater and collection site sediment, and were acclimated to 15° C and held until used for the testing on July 30, 1997. Samples numbered 9, 31, 35, 46, 72, 96, as well as the control sample number 39 were analyzed. Tests were conducted according to USEPA (1993). On the day before test initiation, each 4-L bucket of test sediment was homogenized and 500-g portions were added to fifteen, 1-L glass beakers. Three beakers comprise one replicate for each test sediment and five replicates of each test sediment were performed. The jars were then filled with 500-mL of clean seawater (salinity 28 ± 2 ‰), then aerated overnight with oil-free compressed air at a rate of approximately 150 mL /min. The following day, three clams were transferred to each beaker. Three sub-samples of nine test organisms were taken at the beginning of the test. The length, weight, and wet tissue weight were recorded, and the tissue was frozen for 10 chemical analysis. Any animals dead or not buried in the first 24 hours were replaced. Daily recording of observations, temperature, and aeration occurred. Three times a week representative test chambers were analyzed for pH, salinity, temperature, and dissolved oxygen. Three times a week, the overlying water was 80% renewed with clean seawater. Lighting was provided by overhead fluorescent fixtures with 16 hours of light and 8 hours of dark, daily. The temperature was maintained at 15 ± 1° C throughout the test. After 28 days, the clams were removed from the test sediment, rinsed with clean seawater, and placed in clean collection site sediment for 24 hours. The three beakers comprising the one replicate were combined at this point. This depuration period allows for removal of gut content that could interfere with chemical analysis of tissue. The tissue samples were collected and wet weights recorded before submission for chemical analysis. The percentage survival was also computed at the end of the 28-day exposure. 2.3.3 West Coast Analyses. This text, extracted from Fennell and van Aggelen (1997), describes the biological toxicity test methods used at the Pacific Environmental Science Centre while conducting biological toxicity tests on Eohaustorius washingtonianus, Eohaustorius estuarius, Boccardia proboscidea, Polydora cornuta, and Vibrio fischeri. It should be noted that sediments arrived at the lab frozen and were not used for this study. Replacement sediments for the biological toxicity tests obtained from the East Coast lab were used instead. The following text outlines the results of four sediment biological toxicity tests performed in July and August, 1997: 10-day survival tests using two amphipod species; 14-day growth and survival tests using two polychaete species; the Microtox. solid-phase metabolic-inhibition test; and an echinoid fertilization-inhibition test. 2.3.3.1 Amphipod Toxicity Tests. Amphipod sediment testing was performed using two species of infaunal amphipods, E. washingtonianus and E. estuarius. The sediment samples were numbered 2, 6, 11, 15, 23, 24, 33, 34, 38, 40, 48, 71, 73, 75, 80, 84, 87, and 100, as well as the control sample numbers 29 and 13, respectively. Amphipod test set up in EC Toxicology Laboratory in Moncton, NB. Source: K. Doe E. washingtonianus were field collected at Esquimalt Lagoon, Vancouver Island by Biologica Environmental Services. E. estuarius were field collected at Beaver Creek, Oregon by Northwestern Aquatic Sciences. Amphipods were held in control sediment (i.e., collection site sediment) under continuous light and aeration and were acclimated at 15 ± 1° C over two or three days before test initiation. Static 10-day acute survival tests were performed according to the procedures outlined in Environment Canada (1992a). The control sediment used in these tests was homogenized and wet sieved through a 0.5-mm stainless steel sieve to remove native organisms. Each test sediment sample was homogenized by hand. Five acid-washed 1-L jars were prepared for each control and test sediment. Approximately 175 to 200 g of sediment (to a height of 2 cm) was added to each jar. Each container was then carefully filled with a fresh laboratory supply of sand-filtered seawater from Burrard Inlet, being careful not to disturb the sediment layer. The test containers were aerated and allowed to settle overnight. Twenty randomly selected amphipods were added to each of five replicate jars per sediment. Water quality (temperature, 11 pH, salinity and dissolved oxygen) was monitored periodically throughout the test in replicate A. The biological toxicity tests were conducted in an environmental chamber at 15±1° C under continuous light. At the conclusion of the biological toxicity tests, the total number of emergent amphipods on the sediment surface (or swimming in the water column) of each test container was recorded. The sediments were then wet-sieved through a 0.5-mm stainless steel screen, and total surviving, dead and missing amphipods were recorded. Means and standard deviations were calculated for percent survival and percent emergent1. In addition, 96-h LC50 positive control tests were run concurrently, using various concentrations of the reference toxicant CdCl in seawater, to assess the acceptability of test conditions and amphipod sensitivity in reference to historical performance under the same conditions (including such conditions as darkness and absence of substrate). The LC50 values (and associated 95% confidence limits) for the positive reference toxicant tests were determined using the Environment Canada computer program following Stephan (1977). 2.3.3.2 Polychaete Survival/Growth Tests. Fourteen-day polychaete sediment biological toxicity tests and concurrent 96-h positive control reference toxicant tests with cadmium were performed using B. proboscidea and Polydora cornuta, two species of spionid polychaetes cultured in-house. When juveniles were almost three to three-and-one-half weeks old, they were considered ready for use in toxicity tests. Control sediment consisted of sieved (500 µm) and rinsed (clean natural seawater) sediment from the polychaetes’ natural environment. B. proboscidea sediment was collected from Witty’s Beach, Vancouver Island, BC, and P. cornuta control sediment was collected from Conrad’s Beach, NS. Samples numbered 6, 66, and 68 as well as the control sample numbers 32 and 4 were analyzed. Tests were conducted 1 Not analyzed. according to Environment Canada’s draft protocol (1995a). A controlled environment room was set to uniformly maintain 23 ± 1° C and a photoperiod of 16 hours light to eight hours dark. The test vessels were prepared the day before the polychaete introduction (Day-1). Five acidwashed 1-L glass mason jars were each filled with 175–200 mL (to a height of 2 cm) of a test sediment, to which 750–800 mL of clean control/dilution water was added (fresh laboratory supply of sand-filtered natural seawater from Burrard Inlet). The jars were aerated with filtered, oil-free compressed air overnight, and for the duration of the test, at a steady rate of approximately 150 mL/min/L through plastic aquarium airline tubing and precut, disposable 1-mL polystyrene pipettes. On Day 0 (test initiation day) rearing vessels containing polychaetes of appropriate testing age were sieved. Five polychaetes were added per test replicate chamber. Pre-weighed aluminum pans containing a known number of test age juveniles were dried overnight (60° C) for initial weight determination. Water quality parameters (pH, dissolved oxygen, temperature, and salinity) were measured periodically throughout the 14-d test period. Approximately 80% of the overlying water was replaced on Day 7. Test organisms were fed every Monday, Wednesday, and Friday 500 µL per test vessel (= 5 mg per worm). The food consisted of a 50:50 by weight blend of ground Enteromorpha spp.: Tetramarin. (a green alga and commercial fish flake, respectively), each ground to a fine powder and mixed up into a seawater slurry (2.5 g E:T/50 mL seawater). On Day 14, test vessels were sieved (500 µm) and the numbers of surviving, dead, and missing polychaetes were recorded. Surviving polychaetes were rinsed in de-ionized water before placement in pre-weighed aluminum pans for final dry weight determinations. 12 The means and standard deviations for percent survival and growth achievement of polychaetes exposed to contaminated test sediment samples and negative control sediment were calculated. The negative control (sediment from original polychaete collection site) provides not only a basis for interpreting data obtained from any reference and test sediments, but also provides evidence of the relative quality of the test organisms and suitability of test conditions and procedures. An acceptable (= 90%) control survival level must be achieved for a test to be considered valid. The LC50 values (and associated 95% confidence limits) for the positive control reference toxicant tests were determined using the Environment Canada computer program based on Stephan (1977) and were compared with values derived from previous reference toxicant testing. 2.3.3.3 Marine Photoluminescent Bacterium Acute Toxicity Test. A marine bioluminescent bacterium, Vibrio fischeri, was used to assess the toxicity of the test sediments using the Microtox. test system. Vials of freeze-dried V. fischeri stored at –20 ± 2° C were reconstituted in 1.0 mL of distilled water and incubated at 5.5 ± 1° C for no less than 20 min before use in solid-phase tests. Test results were based on measured light output in the presence of various levels of test substance in aqueous solutions, which were compared with light output of a control blank (i.e., bacterial cell suspension in diluent only). Light output is a product of the electron transport system and relates directly to the metabolic state of the bacteria (Schiewe et al., 1985). The degree of light loss (degree of metabolic inhibition in the bacteria) indicated then the degree of toxicity of the sample. The sediment remaining in one polystyrene 50-mL tube following centrifugation was homogenized before solid-phase testing carried out according to methods outlined by Microbics Corporation (1992). Bacteria were incubated for 20 min at ambient room temperature in a series of aqueous solutions of various concentrations made up of the sediment sample and a 3.5% solution of Reagent Grade NaCl crystals dissolved in de-ionized water. Following this incubation period of direct bacterium-particle interaction, the solutions were filtered and 500 µL of each filtrate was transferred to a corresponding glass cuvette within the incubation unit. After a further five-minute incubation period at 15.0 ± 0.5° C, light emission from each concentration was measured. A Microtox. model 500 Toxicity Analyzer (Beckman Instruments, Carlsbad, CA) controlled by the appropriate Microtox. software (version 7.03) was used for all procedures. A dose-response curve was determined by Microbics software (version 7.03 for solid-phase), on which the IC50 was located. A 95% confidence range was also reported. The IC50 is the inhibiting concentration of a sample causing a 50% decrease in the bacterial light output under defined conditions of exposure time and test temperature. IC50s derived from solid-phase testing were corrected for moisture content by standard laboratory procedures based on Microbics Corporation (1992) using the remaining sediment in the tube on the day of testing and oven-drying (overnight at 100 ± 5° C) three replicates of 5.0 ± 0.2 g per sediment sample. Microtox. Model 500 Toxicity Analyzer. Source: K. Doe 13 2.3.3.4 Echinoid Fertilization Inhibition Assay. Fertilization inhibition tests were performed using the gametes of the echinoderm Dendraster excentricus (eccentric sand dollar). Sand dollars spawned for collection of gametes were field-collected in May 1997 at low tide from Crescent Beach, White Rock and were held at the laboratory in an outside tank with a 7–8 cmbed of Crescent Beach sand and a source of flowing seawater. Testing procedures were those outlined in Environment Canada (1992b) and British Columbia Ministry of Environment, Lands and Parks (1994). Samples numbered 2, 6, 11, 15, 23, 24, 33, 34, 38, 40, 48, 71, 73, 75, 80, 84, 87, and 100 were analyzed. Five full 50-mL polystyrene tubes per test sediment were centrifuged for 30 min at 4000 rpm and 4° C to extract the pore water from the sediment. The interstitial water was collected into beakers and water quality parameters were measured. Control/dilution water was a laboratory supply of sand-filtered seawater from Burrard Inlet which was subsequently filtered through a 0.8-µm filter, adjusted with natural brine salts to match the salinity of the most saline pore water samples within 2‰to a minimum of 28 ‰, aerated gently, and held at 15°C. A positive reference toxicant test using a range of CuSO4 concentrations was run concurrently to measure species sensitivity and acceptability of test conditions. Following wet spawning of the sand dollars, collected sperm and eggs were kept separate to avoid gamete contamination. For each gender, gametes from at least three individuals were pooled and after density determinations, dilutions were made to achieve a final sperm-to-egg ratio of 2000:1 in a 2.0-mL test volume. Initially, sperm were exposed for 10 min to three replicates of full strength (100%) pore water obtained from each sediment sample. Following the 10-min sperm-only exposure, eggs were added for an additional 10-min exposure period. Immediately thereafter, the samples were preserved with 10% buffered formalin to fix the eggs. Fertilization rates in each pore water sample were determined by calculating the average for all replicates of the number of eggs with fertilization membranes counted out of the first 100 eggs encountered under a microscope for each replicate. The results from the positive reference toxicant test were adjusted using Abbott’s formula (Finney, 1971) to correct all values for mean percent unfertilized eggs at test end, in keeping with the variable and gamete-dependent differences from test to test with respect to fertilization success rate and the associated percentage of unfertilized control eggs. Thereafter, Environment Canada’s statistical package for calculating LC50s, based on Stephan (1977) was used to compute the IC50 (and its associated 95% confidence limits). Benthic Macroinvertebrate Survey 2.4 This text was extracted from the report by Arenicola Marine (1997). 2.4.1 Data Manipulation. Data was manipulated as follows: • Taxa with no entries were deleted. The only taxon included in the taxa list that did not have any entries was Tharyx marioni. 2.4.2 Data Presentation. The following indices of benthic community structure were calculated and are presented graphically in Section 3.2, total number of organisms, number of species, Simpson’s Diversity Index, Pielou’s J (evenness), McIntosh’s Index (evenness), Margalef’s Index (diversity), and Shannon’s H (diversity). The total number of organisms and the number of taxa per square metre are summarized for each station. It is generally thought that a community suffering a deleterious impact will be characterized by a relatively small number of organisms belonging to a few taxa. 14 Simpson’s Index (1949) is a diversity index measuring the probability that two organisms chosen at random from a population will belong to the same taxa. A diverse community will produce a low value of Simpson’s Index. This index does not account for the non-uniform distribution typical of benthic macroinvertebrates. The index is also a function of sample size; thus comparison of samples from different locations or samples collected using different methods are confounded by sample size. Shannon’s H´ as described in Shannon and Weaver (1949) is a diversity index founded on information theory. The validity of this measure of diversity is largely a consequence of different interpretations of the term “diversity.” It has been shown that diversity may increase even when the species numbers decrease, if evenness increases (Hurlbert, 1971). Wilhm (1970) suggests that diversity indices > 3 are characteristic of diverse sites, while values < 1 suggest gross pollution. Evenness is a measure of how evenly species are distributed across taxa. Pielou’s J is a measure of evenness and is often given by H´/ H´max. It has a maximum value of 1. Diversity and evenness measures are usually highly correlated. McIntosh’s Index is a measure of equitability or evenness similar in form to that of Pielou’s J with an index being divided by the maximum value attainable by that index. The index is based on the Euclidean measure of distance; consequently, it is also known as McIntosh’s Ecological Distance (McIntosh, 1967). Margalef’s Index (1958) is another diversity index and assumes a linear relationship between species abundance and number of species. 2.5 Sediment Physical/Chemical Analyses 2.5.1 Data Manipulation. • All sediment samples were subsampled to produce two or three pseudoreplicates. One of the two sediment subsamples for PAHs for the St. Ann’s Harbour reference station, was further split into two sub-subsamples. The two sub-subsamples were averaged and then averaged with the single subsample. When present, subsample values were averaged and averages were used in subsequent calculations. • Metal levels below the detection limit were replaced by the detection limit. 2.5.2 Oxidation Reduction (Redox) Potential, Ammonia, and Sulphide. Sediment samples were thoroughly homogenized and subsampled for analysis of sulphide, redox potential (Eh), and ammonia by specific ion electrode according to the manufacturer's instructions and advice by Dr. B. Hargrave (Dept. of Fisheries and Oceans, Scotia–Fundy Region). These analyses were conducted in triplicate. Results for sediments are expressed as µg S/g dry weight of sediment for sulphide, µg NH3-N/g dry weight of sediment for ammonia and millivolts corrected for the normal hydrogen electrode for redox potential. Other subsamples of these sediments were centrifuged at 3000 rpm for 15 min, the porewater was decanted off and analyzed for ammonia and pH. Porewater ammonia is expressed as mg NH3-N/L, the pH is expressed in pH units. Testing was conducted on July 30 to August 1, 1997. Samples numbered 2, 6, 11, 15, 23, 24, 33, 34, 38, 40, 48, 71, 73, 75, 80, 84, 87, and 100 were analyzed. 2.5.3 Sediment Metal Concentrations. Total sediment metal concentrations were measured using inductively coupled plasma atomic emission, graphite furnace atomic absorption, and cold vapour atomic fluorescence for mercury (Hg) (PESC, 1999a). 2.5.4 Porewater Metal Concentrations. Total porewater metal concentrations (with the exception of porewater Hg) were measured using inductively coupled argon plasma atomic emission spectrometry, and graphite furnace atomic absorption spectrometry (PESC, 1999b). Porewater Hg was measured using cold vapour 15 atomic fluorescence spectrometry following acid digestion (PESC, 1999c). 2.5.5 Sediment Organic Compound Concentrations. Sediment organic compounds (PAHs and PCBs) were measured by the Atlantic Region Monitoring and Evaluation Branch (EC, 1997) method. This involves extraction into 1:1 mixture of hexane and ethane while sonicating. The extract is mixed with acidified water, and back-extracted using hexane. The second stage extract is dried using anhydrous sodium sulphate, and cleaned with a silica-gel mini-column. Further cleanup using toluene follows. The extract is made up to volume and is analyzed using gas chromatography with mass spectrometric detection for PAHs, and gas chromatography with electron capture detection for PCBs. 2.5.6 Porewater Organic Compound Concentrations. The analysis of PCBs and PAHs was conducted following methods outlined in the Atlantic Region Environmental Quality Laboratories (EC, 1992). Porewater was extracted by centrifugation (Jackman and Doe, 1997). Organic compounds are extracted into hexane, dried by passing through anhydrous sodium sulphate, then cleaned if necessary on a silica-gel mini-column. The extract is made up to volume and analyzed using gas chromatography with mass spectrometric detection for PAHs, and gas chromatography with electron capture detection for organochlorine compounds, PCBs, and chlorinated benzenes. 2.5.7 AVS and SEM Metals. The extraction procedure for AVS and SEM metals follows Allen et al., (1993) and was performed by Dundas Environmental Services, Burlington, Ontario. Unfortunately, SEM Ni was not measured by the contractor and is therefore not included in the estimation of total SEM. 2.6 Tests of Hypotheses The methods used to test each hypothesis are presented in Section 4 “Tests of Hypotheses.” Section 3 Biological Toxicity Tests %Survival Survey Observations and Biological Toxicity Test Results 3.1 The results of biological toxicity tests are best presented using Box and Whisker plots. Due to the lack of this type of plot in Microsoft Excel®, a similar plot is presented without the inclusion of boxes. These graphics show the maximum, minimum, median and 25th and 75th percentiles. 100 90 80 70 60 50 40 30 20 10 0 Control Figure 2 Percent Survival for Amphiporeia virginiana. Survival is very low at Station 1, but gradually improves along the gradient. Note that survival in the reference site is still not as 9 6 5 1 Stations high as the control sediment. The following plots use the subsamples as raw data. Thus the sample size for a site with three field replicates and five lab replicates or subsamples, is 15. This procedure is used to present the raw data and is not used during data analyses in Section 4. 3.1.1 Acute Tests for Sediment Toxicity Using Marine Amphipods. Results from these tests are shown in (Figures 2, 3, 4, and 5). 12 St. Ann's Harbour 16 Minimum 25th Percentile Median 75th Percentile Maximum 100 % Survival 100 90 80 70 60 50 40 30 20 10 0 Figure 3 Percent Survival for Rhepoxynius abronius. A minimum survivorship is seen at Station 5 with a gradual improvement in survivorship to the reference station. This response is not as 90 80 70 60 50 40 30 20 10 0 Figure 4 Percent Survival for Eohaustorius washingtonianus. The survival of Eohaustorius %Survival 1 Control strong as that observed for the other amphipods and is obscured by variability. Control 1 5 6 9 12 St.Ann's washingtonianus is similar to that of the other amphipods. 17 Minimum 25th Percentile Median 75th Percentile Maximum 12 9 6 5 St. Ann's Harbour Stations Minimum 25th Percentile Median 75th Percentile Maximum Harbour Stations 18 100 90 80 70 60 50 40 Minimum 25th Percentile Median 75th Percentile Maximum 30 20 10 0 Control 1 5 6 9 12 St.Ann's Harbour Stations Figure 5 Percent Survival for Eohaustorius estuarius. Again, there is a decreased survivorship at Station 1 with a gradual increase in survivorship to the reference station. 3.1.2 Sublethal Toxicity Test for Metabolic Inhibition Using a Marine Bacterium. moisture-corrected basis. Moisture-corrected IC50s are presented in Figure 6. The Microtox. solid-phase toxicity test was performed on both a wet-weight and Minimum 25th Percentile Median 75th Percentile Maximum 20000.00 18000.00 16000.00 14000.00 12000.00 10000.00 8000.00 6000.00 4000.00 2000.00 0.00 12 9 6 5 1 St. Ann's Harbour Stations Figure 6 Photoluminescent Bacteria Light Inhibition. The Microtox. assay shows an increase in IC50 (ppm moisture corrected) %Survival IC50 along the gradient with greatly elevated IC50 at the Station 12 reference site but a decrease at the St. Ann’s Harbour reference site. The mean IC50 for stations, 1, 5, 6, 9, 12 and St. Ann’s Harbour are 97, 122.67, 144.67, 1009.33, 13200, and 1733.333 ppm, respectively. 3.1.3 Sublethal Test for Sediment Toxicity Using Marine Polychaetes. The results of toxicity tests conducted by the East coast laboratory using Polydora cornuta and Boccardia proboscidea are presented in Figures 7, 8, 9, and 10. Polychaete 100 90 80 70 60 50 40 30 20 10 0 Figure 7 Percent Survival for Polydora cornuta. The survivorship for Polydora cornuta is reduced at Station 9 and is extremely variable at Stations 9 and 12. Note that approximately half (17/35) of the tests showed complete survival. 6 5 4 3 2 1 0 Figure 8 Growth for Polydora cornuta. Growth is depressed at intermediate stations relative to the control station and St. Ann’s Harbour. %Survival Weight / worm (mg) 19 testing was also conducted at some stations along the gradient by the Pacific Environmental Science Centre where sufficient sample was available. These data have not been incorporated into this analysis but are presented in Appendix A; West Coast Polychaete Analyses. Minimum 25thPercentile Median 75thPercentile Maximum Control 1 5 6 9 12 St.Ann's Harbour Stations Minimum 25th Percentile Median 75th Percentile Maximum 12 Control 9 6 5 1 St. Ann's Harbour Stations 100 90 80 70 60 50 40 30 20 10 0 1 Figure 9 Percent Survival for Boccardia proboscidea. As for Polydora cornuta, most organisms have high survival in the sediments. 2.5 2 1.5 1 0.5 0 1 Figure 10 Growth for Boccardia proboscidea. There seems to be no trend in Boccardia proboscidea growth. Weight / worm (mg) %Survival 9 Stations 9 Stations 20 Minimum 25th Percentile Median 75th Percentile Maximum St. Ann's Harbour 12 Minimum 25th Percentile Median 75th Percentile Maximum St. Ann's Harbour 12 21 is presented for Dendraster excentricus (Figure 3.1.4 Sublethal Toxicity Test for Echinoid Fertilization Inhibition. Fertilization inhibition 11) and Lytechinus pictus (Figure 12). Minimum 25th Percentile Median 75th Percentile Maximum 90 80 70 60 50 40 30 20 10 0 12 9 6 5 St. Ann's Harbour %Fertilization % Fertilization 100 1 Control Stations Figure 11 Percent Fertilization for Dendraster excentricus. The percent fertilization decreases along the gradient and is a maximum at the reference station. This is the opposite of what is expected along an organic contaminant gradient. Note that the St. Ann’s harbour station performs similarly to the control sediment whereas the Station 12 reference sediment does not. Note the large degree of variability. 100 Minimum 25th Percentile Median 75th Percentile Maximum 90 80 70 60 50 40 30 20 10 0 Control 1 5 6 9 12 St.Ann's Harbour Stations Figure 12 Percent Fertilization for Lytechinus pictus. The results are generally variable with lowest fertilization rates seen at Station 9. Like the results with Dendraster excentricus the reduced fertilization rate at Station 9 relative to Stations 1, 5, and 6 is unexpected. Note that the St. Ann’s harbour station does not perform as well as the station 12 reference station for Lytechinus pictus. Again, note the large degree of variability. 90 Table 2 Response 100 % Survival 80 70 60 50 40 30 20 10 0 Percent Survival Percent Survival Percent Survival Percent Survival IC50 (ppm moisture V. fischeri corrected) Percent Survival Growth (mg) Percent Survival Growth (mg) Percent Fertilization D. excentricus Percent Fertilization L. pictus 3.1.6 Bioaccumulation Test Using Bivalves. The survival results for the Macoma nasuta bioaccumulation study are presented in Figures 14, 15, 16, and 17. Tissue levels for contaminants of concern are presented below. Figure 14 Percent Survival for Macoma nasuta. Survival is quite variable, but 100% survival occurs in at least one test from every site. Also, decreased survival is Summary of Mean Biological Test Responses Test 5 1 Control 3.00 83.33 52.00 99.33 86.00 72.67 100.00 65.33 85.67 A. virginiana R. abronius E. estuarius E. washingtonianus 97.00 47.67 65.67 79.33 94.67 97.33 94.67 122.67 144.67 1009.33 13200 97 92.00 80.00 96.00 1.03 2.11 P. cornuta P. cornuta B. proboscidea B. proboscidea 0.69 100.00 1.55 93.83 66.11 48.89 80.67 56.50 39.00 Relationships between tissue contaminants and sediment and porewater contaminants are examined in Section 4.2.4. 12 9 Control 6 5 1 Stations observed at Station 9, which is unexpected, given the organic contaminant gradient. 23 Station 6 53.00 81.67 83.67 56.33 84.00 0.95 58.11 61.92 St. Ann's Harbour 12 9 74.00 79.67 96.33 88.33 St. Ann's Harbour 77.00 92.33 96.67 82.33 1733.333 96.00 2.24 93.75 1.58 95.11 12.50 72.00 0.95 87.50 1.41 4.67 58.83 44.00 0.47 87.50 1.56 9.33 9.50 Minimum 25th Percentile Median 75th Percentile Maximum 24 500 450 400 Minimum 25th Percentile Median 75th Percentile Maximum 350 300 250 200 150 100 50 0 Control 1 5 6 9 12 St.Ann's Harbour Stations Figure 15 PCB Tissue Concentrations in Macoma nasuta. PCB tissue levels reach a maximum at Station 5 and decline to control levels at the reference station. Note that as survival at Station 9 was minimal, bioaccumulation at that station was not expected. 10000 9000 Minimum 25th Percentile Median 75th Percentile Maximum 8000 7000 6000 5000 4000 3000 2000 1000 0 12 Control 9 6 5 1 St. Ann's Harbour Stations Figure 16 PAH Tissue Concentrations in Macoma nasuta. PAH tissue concentrations decrease monotonically from Station 1 closely following the trend in sediment PAHs. Concentration (ng/g dry tissue) Concentration (ng/g dry tissue) 3 2.5 2 1.5 1 0.5 0 Summary of Available Biological Responses Figure 20 Benthic Macroinvertebrate Community Structure Indices. The distribution of organisms across taxa is most even at Station 1 and 6 with the other stations showing lower values for Pielou’s evenness. McIntosh’s index, which is another measure of evenness, provides the same interpretation as Pielou’s index. Note that the St. Ann’s reference station shows the lowest evenness using McIntosh’s index. Simpson’s diversity index shows that the St. Ann’s reference station is the least diverse station. Margalef’s and Shannon’s diversity indices also support this interpretation and show that Station 9 (Margalef’s) and Station 6 (Shannon’s) are the most diverse stations. A listing of the available responses is presented in Table 4. The column entitled Obvious Station Index Value 3.3 Response comments on the presence of a visually obvious response. There may be no 27 Pielou's Evenness McIntosh's Index Simpson's Index Margalef's Index Shannon H 1 5 6 9 12 St.Ann's Harbour response, a response but no trend, or a trend in response. The term along gradient is directional and refers to stations in the following order 1, 5, 6, 9, 12, and St. Ann’s Harbour. Thus an increase in response along the gradient is interpreted as an increase in the response from Stations 1 through to St. Ann’s Harbour. Table 4 Summary of Available Biological Responses Test/Species A. virginiana Acute Responses E. estuarius E. washingtonianus R. abronius B. proboscidea P. cornuta M. nasuta Sublethal Responses B. proboscidea P. cornuta D. excentricus L. pictus Microtox® (solidphase, moisture corrected) M. nasuta Bio-accumulation Benthos In Situ Response* * The compositing procedure precludes enumeration of benthic samples. Therefore, in the strict sense of the term, the benthic samples are not synoptically collected. Response Percent survival Percent survival Percent survival Percent survival Percent survival Percent survival Percent survival Growth rate Growth rate Percent fertilization Percent fertilization Change in luminescence Bioaccumulation Abundance Diversity Evenness Richness 28 Obvious Response? Increasing trend. Increasing trend with asymptote at Station 9. Increasing trend, with reference response lower than control response. A quadratic response curve with a minimum at Station 5. Little response overall, decreased survivorship at Station 12 reference site. The response is variable with a minimum occurring at Station 9. Extremely variable with a minimum at Station 9. No trend seen. Growth is elevated at the St. Ann’s Harbour station and depressed at Station 1. Decreasing trend, with large variability among all responses. Variable with minimum percent fertilization at Stations 9 and St. Ann’s Harbour. Peak IC50 found at Station 12 reference sites with IC50s higher at reference stations than innermost stations. A well-defined peak in PCB tissue concentrations at Station 5 declining monotonically to St. Ann’s Harbour. PAH tissue concentrations decline monotonically to St. Ann’s Harbour. As, Cd, and Hg tissue concentrations are invariant with respect to stations while Cr, Cu, Ni, Pb and Zn levels change with location. Peak abundance at Station 9. Intermediate stations most diverse, reference station, least diverse. Generally decreasing. Peak richness at Station 9. Most of the toxicity tests exhibit responses; often trends in the responses are seen along the gradient. Station 9 seems to be a pivotal station for echinoids, P. cornuta and M. nasuta with minimum responses occurring. This is contrasted with maximal richness and abundance of benthic macroinvertebrates at Station 9. The greatest decrease in amphipod survival and bacterial photoluminescence was observed in organisms exposed to sediments collected from Station 1. Species-specific Application Limits for Reference Method (EC, 1998a) Table 5 Acceptable Physicochemical Characteristics of Test Sediment Test Species Porewater salinity (‰) R. abronius Must be 25 to 35 Must be 15 to 35 E. washingtonianus E. estuarius Must be 2 to 35 A. virginiana Must be 15 to 35 Summary of Grain Size Application Limits Table 6 Station 1 5 6 9 12 St. Ann’s Harbour * “Coarse” sediments consist of gravel + percent passing through a size 30 mesh. This is slightly at odds with the definition presented in Table 5, but was used as the Sydney harbour study was conducted before sediment grain size application limits were established. 29 3.4 Validity of Test Sediments and Biological Toxicity Tests 3.4.1 Suitability of Test Methods for Marine and Estuarine Amphipods. Various criteria are being developed to ensure the validity of interpretations using Environment Canada’s reference methods (EC, 1998a) for marine and estuarine amphipods. The limits for physicochemical characteristics as of July 1998 are shown in Table 5. Table 6 summarizes the grain size distribution at the stations sampled during the Sydney harbour study. Percent very coarse-grained (> 1mm) 0 to 100 is acceptable Must be < 25 Must be < 90 0 to 100 is acceptable Coarse* 2.8 2.3 7.4 3.2 1.0 4.4 Sediment Grain Size Percent Fines (< 0.063 mm) Must be < 90 Must be < 80 0 to 100 is acceptable Must be < 90 Sediment Grain Size Fines 72.3 71.5 62.7 80.7 59.4 68.5 Percent Clay (<0.004 mm) Must be < 40 Must be < 20 Must be < 70 Must be < 35 Clay 14.2 12.4 13.2 19.1 5.2 10.5 The only failure of the stations with respect to application limits occurs at Station 9 where the percent fines criteria for E. washingtonianus is very slightly exceeded. As the exceedance is so slight as to be negligible and the percent survival for E. washingtonianus is high at this station, the complete E. washingtonianus data set is used in subsequent analyses (K. Doe, pers. comm., Environment Canada, Moncton, NB, 1999). 3.4.2 Suitability of Test Sediments for Comparison with Reference Site. A sediment should only be compared to a reference site if the mean 10-day survival in reference site sediments is > 80% for R. abronius and E. estuarius, > 75% for E. washingtonianus, and > 70% for A. virginiana. This proviso ensures that a protective benchmark for comparison of biological toxicity test results is used when making ocean disposal decisions. Replicate # 1 for A. virginiana has a mean 10-day survival of 69%. Therefore Replicate # 1 would not pass the criterion for comparison of reference sites to exposure sites. However K. Doe (pers. comm., Environment Canada, Moncton, NB, 1998) and L. Porebski (pers. comm., Environment Canada, Marine Environment Division, Ottawa, ON, 1998) state that the mean of interest when a client submits an application for disposal of dredged materials would be the mean of all samples and subsamples for a given station. In this case, the overall mean survival rate for the St. Ann’s Harbour reference station is 77% and the comparison of exposure sites to this site would be valid according to current criteria. All site Summary of Required Control Survival Proportions (EC, 1998a) Table 7 Biological Toxicity Test Species A. virginiana E. estuarius E. washingtonianus R. abronius All amphipod toxicity tests listed in Table 7 met the criteria for control sediment and reference station survival. 30 means for the Station 12 reference station meet the criteria for minimum survival. 3.4.3 Species Specific Criteria for Validity of a Test. An amphipod toxicity test must meet certain survival criteria in control and reference sediments to ensure that inferences made using test results are valid (Table 7). Station Pass/Fail Status 3.5 This section compares the biological toxicity test responses with a control test response in accordance with the pass/fail criteria generated by Environment Canada personnel. The interim interpretation criteria are summarized in Table 8. The Environment Canada Atlantic Region Toxicology Laboratory interpreted the biological toxicity tests according to the Environment Canada interim interpretation criteria (Environment Canada, 1996). These results are presented in Table 9 alongside analyses conducted by Zajdlik & Associates. Note that only laboratory replication of the polychaete tests was conducted; therefore, comparison among stations is not possible. Comparisons to the control are made using a simulation test (Edwards and Berry, 1987), similar in principal with Dunnett’s test, at the á = 0.05 level using a one-sided test. Subsampling error is incorporated into the overall error variance. However, the Environment Canada Atlantic Region Toxicology Laboratory used subsamples or laboratory replicates as true replicates, artificially decreasing the estimated variance. Minimum Proportion Survival (%) Reference 70 80 75 80 Control 80 90 85 90 Interim Interpretation Criteria Table 8 Test Amphipod Survival* A statistically significant decrease in survival of at least 20% in test sediments as compared to reference sediments, or 30%, when compared to control sediments. Test under development. An IC50 < less than 1000 mg dry solids/L diluent (ppm). Polychaete Growth and Survival Photoluminescent Bacteria* (solid phase) Echinoid Fertilization* Bioaccumulation A statistically significant decrease in fertilization of at least 25% in test sediment pore water as compared to control water. A statistically significant difference in tissue bioaccumulation from control or reference sediment. * Refer to Environment Canada (1996) for criteria to ensure test validity. of the overall error variance. This has the effect of making the test artificially more powerful. In this case, stations that are not truly different from the control may be shown to be statistically different (from the control). This practice, although commonly encountered is not recommended. In the following table, Environment Canada’s (Atlantic Region Toxicology Laboratory) pass/fail decisions are provided along with those made by Zajdlik & Associates. Table 9 shows that the biological responses of organisms exposed to Sydney Harbour sediments were in general accord. Stations 1, 5, and 6 usually failed the Environment Canada pass/fail criteria using the battery of tests, while Station 9 often failed. L. pictus and D. excentricus seem to be the most sensitive species while R. abroniusmay be the least sensitive species. When St. Ann’s Harbour is used as the reference station, D. excentricus is the most sensitive species and the L. pictus and R. abronius tests are the least sensitive. This reversal of relative sensitivity of Lytechinus pictus is due to very low percent fertilization in St. Ann’s Harbour sediments. When Station 12 is used as the reference station, the tests using E. washingtonianus, A. virginiana, and V. fischeri are the most sensitive and D. excentricus and R. abronius are the least sensitive tests. There is excellent concurrence in the pass/fail status of a station when test sediments are compared with either control sediments or 31 Interim Interpretation Criteria reference sediments. The only exceptions occurred when the performance at the St. Ann’s Harbour reference station was poor relative to test sediments as with the L. pictus test. When Station 12 becomes the reference station very low percent fertilization rates for D. excentricus reverses it’s status as the most sensitive test, making it and the R. abronius tests, the least sensitive. Also, although not shown in the table, effect or cutoff criteria were stricter than the statistical test of significance. For example, the use of only the statistical test of significance would have resulted in an increase in the number of stations found to be different from either the control or reference sediments. A discussion regarding the combination of expert judgement and statistical objectivity is provided in Section 6.6. Using current disposal at sea guidance, it should be noted that only one species would have to be used for each test endpoint. Thus passing decisions (using the pass/fail interpretations in Table 9) would have been made as in Table 10, bearing in mind that field replication is not always required (thus each replicate can provide a separate decision). Decisions assume that the species that passed was selected for the test. Decisions also assume (for simplicity) that the bioaccumulation result from replicate 1 would have been the same for other replicates (in practice this test would be done to confirm). Mitigation assumes that the substance can be considered for disposal at sea with special handling. Table 9 a b c d e f Test/Species A. virginiana b,c E. estuarius b,c R. abronius b,c D. excentricus b L. pictus B. proboscidea growth and survival M. nasuta total tissue PAHf Summary of Sediment Toxicity Test Failures Cutoff Valuea (percent or as stated) 53.33/59.33/5 7.00 70.00/77.33/7 6.67 E. washingtonianus c 67.00/74.67/6 2.33 69.33/74.67/7 2.33 68.83/- 20.33/70.11 55.67/33.83/- 12.50 P. cornuta survival P. cornuta growth EC50 <1000 mg/L Photoluminescent Bacteria (Solid Phase)d statistically significant difference The cut-off value is the minimum absolute change in response required in Table 8, applied to the response at the station being considered as a reference station. Thus the first entry of 53.33% = 83.33 - 30% survival in control sediments. The second value of 59.33% corresponds to the cutoff value when Station 12 is used as the basis of comparison and the value of 57.00% represents the cutoff value when St. Ann’s harbour is used as a reference station. Jackman, P. and K.G. Doe. 1997. Toxicity of Sediments from Sydney Harbour, NS - Results of the 1997 Pollution Gradient Study. OR Calculations performed by K. Doe, (pers. comm., Environment Canada, Moncton, NB, 1999) using methods described herein. Pass/fail decision based on a statistically significant decrease in survival of 30% when comparing to control sediment or a 20% when comparing to reference sediment. Format is control sediment cutoff/Station 12 reference site sediment cutoff/St. Ann’s Harbour reference sediment cutoff. Mean moisture corrected IC50s for each station compared to pass/fail criteria of 1000 mg dry solids/L diluent (~ ppm). Based on a statistical comparison of subsamples or laboratory replicates from exposure and control sediments (Jackman and Doe, 1997). This comparison uses subsamples rather than replicates; therefore, the comparisons are artificially powerful. Station/Replicate Failing Pass/Fail Criteria (Jackman and Doe, 1997b) 1/1,2,3, 5/3, 6/2,3 1/1,2,3 1/1,2,3, 5/1,3, 6/1,2,3 5/1 1/1, 5/1,3, 6/1,3, 9/1,2,3, 12/1,2,3 1/1, 5/2, 5/3, 6/2, 9/1, 9/2, 9/3, 12/3, St. Ann’s Harbour/1, St. Ann’s Harbour/2, St. Ann’s Harbour/3 Nonee Nonee 1, 9e 1/1,2,3, 5/1,2,3, 6/1,2,3, 9/1 32 Station Failing Pass/Fail Criteria with Respect to: Control Sediment 1, 5, 6 1 1, 5, 6 None 1, 5, 6, 9, 12 5 , 9, St. Ann’s Harbour No replication. No replication. No replication. No replication. No replication. No replication. No replication. No replication. No replication. 1, 5, 6 1, 5, 6 but no replication St. Ann’s Harbour Station 12 reference 1, 5, 6 1, 5, 6 1 1 1, 6 1, 5, 6 None None 1, 5, 6, 9, 12 None None 5, 9 1, 5, 6 1, 5, 6 1, 5, 6 but no replication 1, 5, 6 but no replication Table 10 Theoretical Pass/Fail Decisions by Field Replicate* Replicate 1 5 6 1 All amphipods - F All echinoids - F Microtox® - F Bioaccumulation - F 2 Repoxinius - P All echinoids - P Microtox® -F 3 Repoxinius - P All echinoids - P Microtox® - F DECISION No disposal at 1 No disposal at 2 No disposal at 3 * Only one sample was collected per station for bioaccmulation tests. Therefore the level of replication is on the subsample level. The statisical comparison uses subsamples rather than replicates; therefore, the comparisons are artificially powerful. The pass/fail decision for bioaccumulation is based on the same data set. The bioaccumulation pass/fail decision is only presented in the first row of Table 10. A. virgin & E. est - P A. virgin, R. abron. & E. est - P L. pictus - P L. pictus - P Microtox® - F Bioaccumulation Microtox® - F Bioaccumulation - F - F All amphipods - R. abron. & E. est - P D.excentric. - P P D. exentric.- P Microtox® - F Microtox® - F R. abron. & E.est- P R. abron & E. est - P L. pictus - P Microtox® - F All echinoids - F Microtox® - F No disposal at 1 No disposal at 1 No disposal at 2 No disposal at 2 No disposal at 3 No disposal at 3 binding capacity of the porewater. Ni was inadvertently omitted from the SEM analysis and thus the total SEM has been underestimated. However as the measured porewater Ni is below the detection limit of 2 mg/L, it is anticipated that the true SEM/AVS ratios will remain below 1. PCBs are only detectable in the first three stations along the gradient while PAHs are detected in only the first four stations. 3.6.1 Summary of Available Physical/Chemical Measurements (Table 13). Station depths, salinity, temperature, conductivity, and dissolved oxygen concentration of overlying water in increments of 3 m from the surface to the bottom, global positioning system coordinates, and a qualitative description of the sediments collected are also available (see Appendix A; Field Data). 3.6 Sediment Physical/Chemical Analyses The average results for the measured sediment and porewater physical/chemical analyses are presented in this section. Table 11 summarizes the sediment-related variables. The cell format is mean/standard deviation. Table 12 summarizes the porewater variables. Where field replicate data is available, the standard deviation is also presented using the format “mean/standard deviation.” Metals are generally below the detection limit with the exception of Hg at Station 1 and Zn at St. Ann’s Harbour. SEM/AVS ratios never exceed 1 indicating that porewater metals are not in sufficient quantity to exceed the AVS 33 9 12 All amphipods - All amphipods - P P L. pictus - P All echinoids - F Microtox® - P Microtox® - F Bioaccumulation - Bioaccumulation P - P All amphipods - All amphipods - P P L. pictus - P All echinoids - F Microtox® - P Microtox® -P All amphipods - All amphipods - P P All echinoids - F All echinoids - F Microtox® - P Microtox® -P No disposal at 1 Disposal at 1 Mitigation at 2 Disposal at 2 Mitigation at 3 Mitigation at 3 34 Table 11 Summary of Sediment-related Variables Stations Variables 1 5 6 9 H ’s 12 arbour St.Ann Metals (µg/g dry weight) 0.055/0.001 0.024/0.006 0.038/0.001 0.333/0.01 0.486/0.03 0.708/0.305* Hg 15.667/5.132 9.667/0.577 10/0 32.667/2.517 39.333/3.786 41/3 As 0.247/0.012 0.08/0 0.147/0.081 0.467/0.141 0.927/0.085 1.167/0.153 Cd 41.933/2.916 24.1/1.153 33.9/3.534 61.7/5.789 86.567/22.938 81.233/7.217 Cr 37.3/0.4 13/1.732 22.7/0.361 53.7/4.158 73.667/4.637 101.333/3.512 Cu 37/2.646 21/1.732 32/2.646 133.333/5.774 214/7.55 285.667/26.312 Pb 27.333/0.577 18/1 24.667/1.155 37.667/1.155 37/2 34.667/0.577 Ni 2/0 2/0 2/0 2/0 2/0 2/0 Ag 84.233/0.351 56.2/3.812 91.1/1.54 281.567/19.798 865.667/749.436 516.267/50.816 Zn PCBs (ng/g dry weight) -30.847/120.013 -63.561/1.116 -69.131/4.049 642.735/68.498 1186.23/348.893 2095.023/861.506 Total PCBs 4.841/0.314 7.792/0.158 7.202/0.426 5.197/0.365 4.834/0.291 4.853/0.423 DryWt (g) PAHs (ng/g dry weight) 4.785/1.021 21.797/13.866 59.352/108.958 3982.435/886236.982 1617.295/228888.665 599.752/3750.125 Naphthalene 5.56/0.257 39.795/5.591 88.167/262.524 361.33/2907.549 710.909/22090.281 1161.378/84039.966 2-Methyl-Naphthalene 6.363/0.361 29.578/1.222 67.07/143.448 260.151/1466.355 485.867/6977.545 735.118/28447.895 1-Methyl-Naphthalene 0/0 27.785/3.608 56.558/93.875 238.984/989.323 449.565/5786.11 2,6-Di-Methyl-Naphthalene 716.182/24587.576 0/0 1.13/0.179 7.209/2.019 202.436/2349.982 372.337/2627.082 690.127/22186.793 Acenaphthylene 0.868/0.007 2.671/0.094 27.619/148.56 152.429/1997.504 259.197/3569.849 419.286/9481.441 Acenaphthene 0/NA NA NA NA NA NA 2,3,5-Tri-Methyl- Naphthalene 5.144/0.199 12.413/0.389 50.378/135.725 432.14/484.075 901.975/42009.474 1636.157/83833.504 Fluorene 13.347/3.517 2427.151/61099.72 340.478/11139.88 50.542/196.316 4929.827/953506.5 8839.264/2408925 Phenanthrene 9.373/1.342 41.734/7.663 1362.971/32028.61 125.853/1312.1 2787.591/312358 5498.471/1204490 Anthracene 0.979/2.874 15.919/3.593 58.653/121.688 264.325/6701.992 567.783/6774.507 932.712/6463.562 1-Methyl-Phenanthrene 31.941/4.269 2950.918/41360.58 337.512/7905.176 41.23/262.417 6054.457/942833.6 13651.301/4564063 Fluoranthene 24.558/7.272 34.936/51.939 3450.766/25696.41 289.418/5033.69 19589.421/22524310 6888.913/945672.3 Pyrene 22.432/43.278 3147.776/63196.41 283.823/3576.512 24.931/3.463 6848.022/1050705 14229.563/5114055 Benzo(a)Anthracene 23.002/32.062 17391.319/8247566.82 8184.197/1500450.97 3798.197/51278.354 312.285/2220.575 30.772/17.565 Chrysene 37.435/1.487 3476.122/27101.88 261.862/1604.273 19.882/2.329 27017.223/31935430 9470.852/819391.1 Benzo(b)Fluoranthene 8.948/9.41 4.979/0.32 81.708/184.096 1621.6/187614.8 4222.973/167416.1 10760.32/4323773 Benzo(k)Fluoranthene 35 Stations Variables 1 5 6 9 H ’s 12 arbour St.Ann 17.59/2.361 9.162/1.318 2005.741/21176.32 108.579/288.475 5073.646/684284.1 12006.685/5277013 Benzo(e)Pyrene 17.88/2.015 193.527/1247.788 11.375/2.342 3352.986/47869.2 23464.113/24031510 8502.341/1219261 Benzo(a)Pyrene 65.646/40.565 8.977/3.872 1529.271/8079.207 84.455/373.671 4158.546/470902.1 10804.859/4310743 Perylene 40.749/14.372 9.295/5.634 2577.197/49100.06 178.978/736.443 20076.987/17091900 7092.662/38102.6 Indeno(123-cd)Pyrene 0/0 0/0 44.725/75.597 671.267/2868.767 4223.435/620459.431 1871.473/4829.516 Dibenzo(a,h)Anthracene 29.818/6.151 7.296/10.961 2030.201/14800.92 121.207/835.753 5476.129/37894.87 14390.661/8315889 Benzo(g,h,i)Perylene 373.617/834.271 212217.02/1725983000 86926.557/101934100 36918.038/3800310 3179.414/336709.3 446.198/3533.43 Total.PAHs 45.638/56.943 13.573/3.154 32.568/21.234 38.165/52.862 36.828/3.139 NH3 (mg NH3–N/g dry wt.) 43.083/196.521 116.956/3950.251 27.267/61.652 18.009/38.957 61.146/525.56 103.817/1770.73 95.776/1978.824 Sulphide (mg S/g dry wt.) 48.356/32589.205 78.433/7215.678 -21.067/2437.208 -41.044/139.03 -64.322/3836.587 -78.867/694.86 Redox Potential (mV) 75.167/0.961 26.733/1.102 37.133/0.503 68.067/2.203 69.733/1.582 69.1/1.1 Moisture % 35933.33/1305.118 8060/1338.768 18800/1126.943 47900/1808.314 81200/14988.996 96066.67/3544.479 TOC (µg/g) 0.1/0 0.1/0 0.1/0 0.1/0 0.1/0 0.1/0 Gravel % 31.5/0.656 39.8/1.375 18.433/0.586 37.1/3.005 28.5/1.493 27.7/1.609 Sand % 58/2.858 54.2/1.493 61.6/0.854 49.533/3.007 59.133/3.266 58.1/1.664 Silt % 10.533/2.344 5.2/0.625 19.133/1.258 13.233/3.98 12.4/2.563 14.233/0.493 Clay % * The cell format is mean/standard deviation. Metals generally decrease along the gradient but rise again at the reference station although metal levels at St. Ann’s Harbour are not as high as Station 1. PCBs and PAHs generally decrease monotonically along the gradient. Table 12 PCBs (µg/L) PCB PAHs (µg/L) Summary of Porewater-related Variables Station Ammonia (NH3–N mg/L) pH Metals (mg/L, except Hg as µg/L ) Hg As Cd Cr Cu Pb Ni Ag Zn SEM/AVS variables (µmol/g dry) SEM-Cu SEM-Zn SEM-Pb SEM-Cd SEM-Hg Total SEM(µmol/ dry)* AVS (µmol/g dry) SEM/AVS Ratio < 0.33 0.03 < 0.01 < 0.01 < 0.01 < 0.04 < 0.02 < 0.02 < 0.01 0.01 < 0.02 < 0.01 0.04 0.01 0.01 < 0.01 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 * SEM Ni was not measured by the contractor and is therefore not included in the estimation of total SEM. However the porewater Ni values for total extractable Ni is below the detection limit of 2 mg/L. Total PAH Naphthalene 2-Methyl-Naphthalene 1-Methyl-Naphthalene 2,6-Di-Methyl-Naphthalene Acenaphthylene Acenaphthene 2,3,5-Tri-Methyl-Naphthalene Fluorene Phenanthrene Anthracene 1-Methyl-Phenanthrene Fluoranthene Pyrene Benzo(a)Anthracene Chrysene Benzo(b)Fluoranthene Benzo(k)Fluoranthene Benzo(e)Pyrene Benzo(a)Pyrene Perylene Indeno(123-cd)Pyrene Dibenzo(a,h)Anthracene Ben(g,h,i)Perylene 1 8.732/6.93 7.66/0.001 0.121 < 6 < 0.6 0.7 < 0.6 < 6 < 2 < 0.05 < 6 < 0.6 0.6 < 0.6 < 6 < 2 1 2 na 1 2 1 < 0.2 < 0.2 0.131 5.11 0.757 9.29 0.020 15.309 36.04 0.42 0.76 21.11 0.04 0.02 0.02 < 0.01 < 0.03 < 0.02 < 0.02 0.03 0.23 0.33 < 0.02 0.62 2.97 0.70 1.03 3.31 1.28 1.73 2.99 0.74 2.53 0.57 1.86 5 7.897/0.996 7.663/0.004 0.06 < 6 < 0.6 < 0.6 < 0.6 < 10 < 2 < 0.2 0.110 4.53 0.022 8.57 0.051 13.283 34.69 0.38 0.97 11.16 0.03 0.02 0.02 < 0.01 < 0.03 < 0.02 < 0.02 0.02 0.17 0.26 < 0.02 0.28 1.35 0.30 0.56 1.82 0.59 0.98 1.56 0.22 1.50 0.25 1.13 36 6 8.779/2.006 7.697/0 na na na na na na na na 0.079 2.88 0.180 4.51 0.049 7.697 26.77 0.29 0.46 6.10 0.04 0.02 0.02 < 0.01 < 0.05 < 0.02 < 0.02 0.02 0.12 0.16 < 0.03 0.19 0.77 0.22 0.30 0.94 0.31 0.42 0.80 0.13 0.79 0.15 0.57 9 27.617/4.72 7.723/0.022 0.138 0.603 0.032 1.04 0.012 1.825 13.95 0.13 < 0.48 1.50 0.03 < 0.01 < 0.01 < 0.01 < 0.05 < 0.02 < 0.02 < 0.01 0.04 0.05 < 0.01 0.11 0.10 0.09 0.08 0.16 0.06 0.09 0.11 0.03 0.16 0.04 0.21 St. Ann’s Harbour 7.401/0.365 7.46/0.006 12 25.561/6.93 7.833/0.036 < 0.05 < 6 < 0.6 1.3 < 0.6 < 6 < 2 0.06 < 6 < 0.6 < 0.6 < 0.6 < 6 < 2 2.1 < 0.2 0.100 0.572 0.057 2.13 0.026 2.882 23.6 0.12 0.066 0.255 1.65 0.42 0.008 2.401 4.40 0.55 < 0.44 < 0.51 < 0.71 0.02 < 0.01 < 0.01 < 0.01 < 0.04 < 0.02 < 0.02 < 0.01 0.01 0.02 < 0.01 0.05 0.03 0.02 0.02 0.06 0.02 0.04 0.03 0.09 0.07 0.01 0.10 Table 13 Summary of Available Physical/Chemical Measurements Measurement Type SedimentMetals Sediment PCBs Sediment PAHs Sediment Sulphide, Ammonia Sediment Particle Sizes Sediment Total Organic Carbon Porewater Metals Porewater PCBs Porewater PAHs Porewater Ammonia Acid-Volatile Sulphides Simultaneously Extractable Metals Redox Potential Porewater pH Porewater Salinity 37 Number of Field Replicates/Station 3 (Station 12 reference site, missing rep 1, Cd) 3 3 3 3 3 1 (Station 6 data missing) 1 1 3 1 1 3 3 3 Section 4 Tests of Hypotheses During the analyses, it became apparent that the reference station at St. Ann’s Harbour was not behaving as a reference station with respect to the in-situ benthic community structure. As the interpretation of the relative performance of the sediment evaluation tools depends, in part, on how a tool performs relative to a reference station, the definition of the reference condition is discussed. Many definitions for a reference station have been established, but possibly the most pragmatic is: A reference condition is identical as far as possible, to the exposure condition except for the intervention being considered. In the Sydney Harbour study, the intervention of interest was the contaminant gradient, which was primarily PAHs and PCBs. Thus, St. Ann’s Harbour was chosen as a reference station as there were no known PAH or PCB inputs other than atmospheric deposition. Attempts were also made to ensure that non-contaminant factors such as grain size and TOC were similar between the reference and exposure conditions. Even though St. Ann’s Harbour met the preceding definition of a reference condition, data analyses showed that the in-situ benthic community there was similar to the community found at the top (most contaminated part) of the gradient. Additionally, Arenicola Marine (1997) concluded that the in-situ benthic community at the St. Ann’s Harbour station was degraded. Thus, evaluation of the relative performance of the sediment assessment tools is made using two different reference sites (i.e., St. Ann’s Harbour and Station 12 reference site). The dual evaluations begin at hypothesis 4. The following sections describe the methods used to test each hypothesis and the conclusions reached. 38 4.1 H1: Homogeneity of Confounding Factors This section addresses the null hypothesis Ho: Are potential confounding factors homogeneous across the stations? This is tested to verify that the choice of stations achieved the study design goal of minimizing the effect of known confounding factors such as TOC, particle size, ammonia, and Eh. The hypothesis is first tested using the nonparametric Kruskal-Wallis test (Tables 14, 15, and 16). Variables that are significantly different among stations (p-value < 0.05) are highlighted. The analysis shows that all variables in Table 14, except porewater pH, vary significantly among the stations. As there is some concern regarding the choice of reference stations, the hypothesis of homogeneity of confounding variables is retested, by separately removing the St. Ann’s Harbour and Station 12 reference sites and re-testing the null hypothesis: Are potential confounding factors homogeneous across the stations? All variables in Table 15, except porewater pH, vary significantly among the stations. Similarly, in Table 16, all variables, except porewater pH and percent clay, vary significantly among the stations. The combined analyses show that porewater pH remains constant among stations. When the Station 12 reference site is omitted, the percent clay composition does not vary significantly among stations, implying that the Station 12 reference site has a significantly different clay composition than other stations. Due to their importance as potential confounding factors, the TOC, Eh, grain sizes, pH, ammonia, and sulphide levels are examined in more detail. The potential confounding variables across the gradient are presented in graphically in Figures 21–26. 39 Table 14 Summary of Kruskal-Wallis Tests: All Stations 6 Variable Kruskal-Wallis ÷2 Degrees of Freedom P-value TOC 16.2156 5 0.0063 Percent Moisture 15.0331 5 0.0102 Sand 15.8175 5 0.0074 Silt 13.4174 5 0.0198 Clay 13.6866 5 0.0177 Eh 30.5927 6 < 0.0001 Sediment Ammonia 40.7095 6 < 0.0001 Porewater Ammonia 46.6413 6 < 0.0001 Porewater pH 11.1025 6 0.0853 Sulphide 47.4889 < 0.0001 Table 15 Summary of Kruskal-Wallis Tests: St. Ann’s Harbour Omitted 42.149 Variable Kruskal-Wallis ÷2 Degrees of Freedom P-value TOC 12.9815 4 0.0114 Percent Moisture 11.2952 4 0.0234 Sand 12.4138 4 0.0145 Silt 11.9296 4 0.0179 Clay 11.1449 4 0.025 Eh 32.4888 5 < 0.0001 Sediment Ammonia 34.7565 5 < 0.0001 Porewater Ammonia 41.0016 5 < 0.0001 Porewater pH 5.2082 5 0.391 Sulphide 5 < 0.0001 Table 16 Summary of Kruskal-Wallis Tests: Station 12 Reference Site Omitted 5 Variable Kruskal-Wallis ÷2 Degrees of Freedom P-value TOC 12.9815 4 0.0114 Percent Moisture 11.2952 4 0.0234 Sand 12.8813 4 0.0119 Silt 9.5253 4 0.0492 Clay 9.3751 4 0.0524 Eh 21.9077 5 0.0005 Sediment Ammonia 25.0282 5 0.0001 Porewater Ammonia 34.4793 5 < 0.0001 Porewater pH 8.8385 5 0.1157 Sulphide 36.0984 < 0.0001 Generally, the St. Ann’s Harbour station exhibits characteristics (TOC, sediment, and porewater NH3–N, and sulphide) similar to stations at the top of the gradient. However, porewater pH is markedly lower at the St. Ann’s Harbour station, although this was not statistically significant. 4.2 H2: Predicting Biological Toxicity Test Responses This section addresses the following null hypotheses: • Ho 2a: Do confounding factors affect the biological toxicity test response? and, • Ho 2b: Are the dose responses predictable? The previous section has shown that some of the confounding factors vary significantly across stations. Thus, differences in responses may be due to the PAH gradient and/or confounding factors. The relationship between the independent variables and the nine biological toxicity tests is now examined. 4.2.1 Correlation Between Independent Variables. The correlation between toxicants, sediment physical/chemical characteristics among stations is examined to provide information regarding how these parameters covary. This is extremely important in model building and interpretation. An example follows. Consider two sediment variables, [Cu] and percent Clay. If a very high correlation exists between these two variables, then a model such asMortality = a + b * [Cu], is functionally the same as Mortality = a + b * percent Clay, i.e., both models will produce similar estimates of mortality. A stepwise model building algorithm will choose either [Cu] or percent Clay as an independent variable, according to various criteria, but not both, due to their high correlation. Depending on the criteria used in the statistical package, either model could be produced. 43 43 40 The interpretation of the two models, however, varies considerably. In one case, it could be said that the toxicant does not affect mortality or alternatively, the response is explicable by environmental variables. The other model states that the observed mortality is related to sediment [Cu]. Thus if the correlation between the independent variables is known, a priori, contending models can be better interpreted. Appendix A; Table A-19 contains the rank correlation (r) between pairs of sediment contaminants whose absolute value is greater than an arbitrary value of 0.90. From this table we see that the PAHs are highly correlated with one another. This degree of correlation implies that there is a large amount of redundant information in the data set. This type of correlation makes the model fitting and evaluation process difficult. We proceed by substituting total PAHs for the individual PAHs. Table 17 contains the rank correlation (r) between pairs of independent variables (with individual PAHs removed) whose absolute value is greater than an arbitrary value of 0.90. Note that other pairs of variables are also correlated but not as strongly. The metals here are highly correlated with one another. As there is currently no useful generic summary for metals the following sediment variable data set is used in subsequent analyses: As, Cd, Cr, Cu, dry weight (of sediment) Hg, Ni, PCBs, Pb, Total PAHs, and Zn. The correlation between porewater variables is investigated. As porewater metals, PAHs, and PCBs were measured only once per station while other porewater variables were replicated, averages of the replications are used in the estimation of the correlations. Table A-20 (Appendix A) contains the rank correlations (r) between pairs of independent porewater variables with absolute values greater than an arbitrary value of 0.90. In this table, as with the sediment variables, the porewater PAHs are highly correlated. Total PAHs are used to summarize the PAH data set and re-estimate the correlations among the porewater variables. Table 18 contains the rank Table 17 Summary of Rank Correlations among Reduced Sediment Variables with |r|0.90 Variable 1 As As As Cd Cd Cd Cu Cu Cu Hg PCB Table 18 Summary of Rank Correlations among Reduced Porewater Variables with |r|0.90 Variable 1 Cr Cr PCB PCB Zn Ammonia correlations (r) between pairs of independent variables (with individual PAHs removed) with absolute values greater than an arbitrary value of 0.90. There is a marked decrease in the number of correlated variables when PAHs are summarized as total PAHs. The following regression model building procedures will use total porewater PAHs as a surrogate for individual porewater PAHs. Note that a negative sign preceding the estimated correlation indicates an inverse correlation. For example, when ammonia 44 45 Variable 2 Cd Cu Pb Cu PCB Pb Hg PCB Pb Pb Pb Variable 2 Zn pH Eh PAH pH Sulphide increases, sulphide decreases. The magnitude of the absolute value of the correlation, determines how strong the relationship is. Correlations close to 1 or –1 are strong, those close to zero, are weak. 4.2.2 Automated Model Building for Biological Toxicity Test Responses. The responses on the independent variables are regressed using an automated model building approach due to the large number of potential independent variables (12 porewater and 12 sediment). Note that automated model building Rank Correlation 0.9074 0.9254 0.9392 0.9608 0.9577 0.9698 0.9071 0.9203 0.9736 0.9326 0.9345 Rank Correlation 0.9899 -0.9837 -0.9187 0.9813 -0.9530 -0.9142 Table 19 Summary of Regression of Biological Toxicity Test Responses on Sediment and Porewater Variables Variable A. virginiana Biological Toxicity Test -0.582 Intercept Ag Ammonia Cr Hg Eh 0.309 Pb PCB pH Salinity Sulphide -0.223 Total PAHs Zn 0.816 Logistic 0.0718 Ammonia As Cd Cr Cu DryWeight Hg Ni PCB Pb Total PAHs Zn Moisture TOC % Sand % Silt % Clay Goodness of Fit* Model Model Pvalue * As the deviance of binomial generalized linear model (GLIM) is only a function of the fitted values and not the observed values, the usual R2 cannot be estimated. Instead, "an index of correlation" or pseudo-R2 is presented for the logistic model. This is the Spearman rank correlation coefficient between observed and predicted values. Survival E. estuarius E. washingto nianus 3.151 -1.290 0.764 -0.131 -0.0434 -0.0639 0.0787 0.907 0.823 Logistic 0.0470 Logistic 0.0275 46 Percent Fertilization D. excentricus R. abronius 397.939 0.854 Porewater Variables -42.416 -273.379 17.510 0.01361 -9.421 Sediment Variables -2.467 0.316 -4.307 -0.167 -1.624 4.817 1.303 -0.00161 0.0404 0.00742 0.366 -0.1390 0.991 0.659 Logistic 1.000 Logistic 0.0231 Weight Gain L. pictus B. proboscidea 1.718 -346.373 -0.155 -1.646 53.834 0.116 1.875 -0.414 12.471 0.00772 -0.135 -0.0135 -0.439 0.000488 -0.454 0.9686 0.942 Gaussian 0.0158 Logistic 0.484 IC50 V. fischeri P. cornuta 14323.090 0.610 8568.085 0.767 -517.984 0.838 0.891 Gaussian <0.00001 Gaussian 0.00466 4.2.3 Consensus Model Building for Biological Toxicity Test Responses. Automated model building procedures are generally frowned upon as previously described. Due to the importance of the relationship between biological toxicity test responses and sediment physical/chemical parameters, models are also fitted manually. The summary of hypothesis tests section discusses the models built using this approach rather than the automated approach. Model building tools include graphical plots of the data, knowledge of sediment physical chemistry, prior experience modeling sediment toxicity test responses, and an understanding of the statistical models being employed. Thus, in some instances an interpretation or model specification may vary depending on the person interpreting the data. All subjective model interpretations are discussed in the model-building procedures that follow. The model building goal was to determine the most important variables contributing to the observed response, rather than prediction. Therefore goodness of fit was sacrificed for a concise model, but not to the point where highly significant variables were omitted from the model. (The goodness of fit can be perfect if one variable is allowed for each observation in the data set. Unfortunately this model is only a restatement of the data and is not very explanatory!) The model building procedures used follow those generally prescribed in statistical textbooks2. A complete description of the model building procedures used is beyond the scope of this document. Interested readers may 2 These include a criterion for the inclusion of variables to the model with a significant increment in explicable variance with a penalty for overparameterization (following the discussion in the previous paragraph), lack of residual structure, which may indicate a sufficient parameterization, and an understanding of how confounding factors and contaminants interact to best choose which contending variables should be included in the model. 47 consult Collett (1991) and McCullagh and Nelder (1989) for an account of model building in the generalized linear model context. General statistical model building is described in Atkinson (1985). 4.2.3.1 Amphiporeia virginiana. A. virginiana survival is well predicted by the model: logit (probability of surviving) = 1.17027 - 0.00001828032 * total sediment PAHs where logit, refers to the logit transformation; log (p/(1-p). A pseudo R2 value of 0.806 as described in the preceding section, is only marginally lower than the R2 value generated by stepwise methods. Other possible single parameter models include porewater Hg, and PAHs, sediment-associated, Cu and Pb as variables. An examination of the relationship between A. virginiana survival rate and non-contaminant variables shows that total organic carbon and redox potential are strongly correlated with survival rates (r = 0.880 and 0.776, respectively). The correlation between TOC and total sediment PAHs is 0.883 and between redox potential and total sediment PAHs is -0.721. Total PAH concentration is related to sediment TOC, as expected, but the relationship between A. virginiana survival and total sediment PAH is stronger than that with TOC. 4.2.3.2 Eohaustorius estuarius. E. estuarius survival is almost as well predicted as A. virginiana survival. The following model best reflects the data set. logit (probability of surviving) = 3.034739 - 0.1188341* total porewater PAHs where logit, refers to the logit transformation; log (p/(1-p). A pseudo R2 value of 0.746 as described in the preceding section, is lower than the R2 value generated by stepwise methods. Other possible single parameter models include porewater Hg and sediment-associated, Cu, Pb, total PAHs, total PCBs, and TOC. Of these, models using porewater PAHs or sedimentassociated Pb or Cu produce almost identical fits. Thus although total porewater PAHs were selected as the variable best describing E. estuarius survival, sediment-associated Pb ( > TEL at 5 of 6 stations and > PEL at 3 of 6 stations) or Cu ( > TEL at 5 of 6 stations, but < PEL at 6 of 6 stations) could equally well have been chosen. Again, the effects of TOC cannot be discounted as it is extremely highly correlated with Pb and Cu ( r = 0.955 and 0.982, respectively). Note the similarity of this model with that generated using automatic model building methods. In both models total porewater PAHs are significantly associated with survival. The manual approach chose TOC as a contending variable while the automatic model building procedure included percent silt. The association between TOC and silt explains the apparent discrepancy between these two models. 4.2.3.3 Eohaustorius washingtonianus. Of the initial variables selected to describe E. washingtonianus survival, all were sedimentassociated metals. This occurs, as E. washingtonianus survival is slightly depressed at St. Ann’s Harbour relative to Stations 9 and 12. This coincides with a general decrease in sediment-associated As, Cu, Hg, and Pb from Stations 1 to 12, with a small increase in St. Ann’s Harbour. The best fitting single variable model was survival as a function of sediment-associated Cu. An examination of the model residuals shows a disturbing spread at intermediate values of Cu. Given that Cu is highly correlated with PAHs, PCBs, and TOC (r = 0.934, 0.973, and 0.990, respectively), it is possible that Cu correlates well with E. washingtonianus survival due to the depression in survival at St. Ann’s Harbour but not as well at the intermediate stations, where organic compounds may be having a greater effect than metals. Models are then tested with Cu and either Eh, porewater PAHs, or TOC. None of these models fit significantly better, nor do they reduce the lack of fit at intermediate values of Cu. Consequently, the following model is adopted: 48 logit (probability of surviving) = 2.449244 - 0.02648029* sediment-associated Cu. This model produces a pseudo R2 value of 0.792 that is lower than the R2 value generated by stepwise methods. However, the current model uses only one variable to explain the response and is more scientifically defensible. Other single variable models that are not significantly different from the Cu-only model include sediment-associated As, Hg, and Pb as independent variables. 4.2.3.4 Rhepoxynius abronius. R. abronius survival is moderately depressed at Station 5 and shows only a very slight dose response. Initial correlation analyses show that the same parameter set responsible for decreased survival in E. washingtonianus (sediment-associated metals) is also correlated with R. abronius survival. However the correlations are approximately 30–40% lower. The best fitting single-variable model is given by: logit (probability of surviving) = 1.924936 - 0.008457369* redox potential. This model fits poorly with the bulk of the model fit ascribable to the intercept. The pseudo R2 value of 0.349 is much lower than the R2 value generated by stepwise methods. 4.2.3.5 Dendraster excentricus. D. excentricus percent fertilization gradually decreases from the top of the gradient to the Station 12 reference site and is very high in St. Ann’s Harbour sediments. The variability of the responses as depicted in Figure 11, is very high at Station 1. The relative magnitude of the within replicate (laboratory replicates) and among replicate (field replicates) variability is tested in Section 4.3. Often an increase in variability is associated with a response to a stressor. An examination of potential independent variables shows no association with contaminants. D. excentricus percent fertilization is strongly correlated with porewater and sediment ammonia, sulphide, moisture, and pH. All of these variables covary (see Table 20). higher than at Stations 1, 5, 6, and 9. A logtransformation reduces the effect of the high IC50 at the Station 12 reference site. The log IC50 is modeled as a function of the sediment physical/chemical parameters. The use of the log transformation implies that the difference in response across independent variables is proportional to changes in the independent variable, rather than additive. As an example, if sediment-associated Ni increases by 1 unit, the log10 (IC50) will decrease by 0.1032 units. However, the IC50 will decrease by 10-0.1032 or 0.7885 units. The following variables are considered for inclusion in the model based upon correlations: Eh, sediment-associated As, Cr, Cu, Hg, Ni, and Pb, and TOC and porewater PCBs. A series of model trials, results in the following model: log10 (IC50) = 5.837466 - 0.1032447 * sedimentassociated Ni. The R2 value of 0.908 is higher than the R2 value generated by stepwise methods and contains fewer parameters. Model residual diagnostics are very good. However, models with sedimentassociated Cr, or Eh fit almost as well due to their high correlation with sediment-associated Table 21 Summary of Model Building for Biological Toxicity Test Responses Response A. virginiana E. estuarius E. washingtonianus R. abronius D. excentricus L. pictus B. proboscidea P. cornuta V. fischeri 10 A strong dose response well predicted by total sediment PAHs but also related to TOC. A strong dose response well predicted by total porewater PAHs, but also related to TOC and TOC-associated variables. Good dose response predicted by sediment-associated Cu, possibly due to increase in concentrations at the reference station. Limited dose response with redox potential weakly correlated with percent survival. The dose-response is not predictable. The highest correlations occur with non-contaminant variables. The dose-response is not predictable. The highest correlations occur with non-contaminant variables. No dose-response is seen for B. proboscidea weight gain. No dose-response but improved growth at the St. Ann’s Harbour station may be due to pH effects. The log-transformed dose response is very well predicted by sediment-associated Ni. 50 Ni. Note that this model considers sedimentassociated Cr as a possible explanatory variable. 4.2.3.10 Summary of Consensus Model Building for Biological Toxicity Test Responses. Table 21 summarizes the major conclusions of the dose-response models. The reader should refer to specific sections for a more detailed interpretation of the modeled dose response. A strong response to the PAH gradients occurred among the amphipods A. virginiana and E. estuarius. Other organisms responded to different contaminants such as metals. Other than redox potential, and in one instance pH, non-contaminant effects such as grain sizes were not observed. The TOC was usually a contending explanatory variable due to its high correlation with contaminants. The exploration of the various models not presented, shows that non-contaminant variables are also associated with the responses, but not as strongly as with contaminants. The correlation between contaminants and non-contaminant factors, particularly TOC and redox potential was often high. Comments The consensus model building approach produces simpler, more interpretable models than the automated model building approach. These simpler models usually do not fit quite as well as the computer-generated models for reasons described at the beginning of Section 4.2.3. 4.2.4 Model Building for Tissue Bioaccumulation. The same model building techniques are employed to investigate which sediment physical chemical variables are correlated with tissue toxicant levels in M. nasuta. Only the following tissue contaminants showed evidence of change and were further examined: Cr, Cu, Ni, Pb, Zn, total PAHs, and PCBs. Models summarizing a significant proportion of variability in the data set are highlighted in the row “Model P-value.” The table is in the same format as Table 19 and is read in the same manner. The tissue PCBmodel fits the observed data moderately well with an R 2 value of 0.687 and describes a significant proportion of the total variability in the data set as the model p-value of 0.0415 is less than 0.05. The automated model building procedure for tissue PAHs chose TOC as the independent variable that best explains tissue levels of PAHs. However an examination of the correlation between TOC and sediment and porewater levels of PAHs found that the correlations were 0.899 and 0.938, respectively. Consequently, a model was fit with porewater PAHs as the only predictor variable. This model has an R2 of 0.904. The associated F-statistic is only slightly smaller than that of the model in Table 22. There is interest in describing the relationship between environmental contaminants and tissue contaminants, and the contending models are equally acceptable; therefore, the following model for tissue levels of PAHs is adopted: Tissue PAH = 601.3638 + 407.736* Porewater PAH This is an example of how automated model building procedures may produce models that do not incorporate a mechanistic understanding of the process(es) involved. 51 Only the tissue PAH and tissue PCB models describe the variability in the data set to a significant extent. Both of these tissue contaminants are strongly correlated with its porewater concentration. 4.3 H3: Do biological toxicity tests perform consistently within sites? This hypothesis may be assessed using replicates within stations to examine the variability of a given response at a station. Significant differences in subsample variability among the replicates may indicate an inconsistent test, or micro-scale differences in sediment physical/chemical quality, possibly due to poor homogenization techniques. This test of hypothesis is similar in intent, to the control charts using reference toxicants for biological toxicity tests; namely consistency of the test under conditions that are as similar as possible. The test is conducted by partitioning each treatment sums of squares into an among-sites sum of squares and among subsample (or laboratory replicate) sums of squares. The within site sums of squares represents the subsampling error. If a test is excessively variable then the subsampling error will be greater than the treatment error. This is tested using the ratio of the two mean square errors. In tests where the response is survival, logistic regression is used to estimate the two error components. The deviances corresponding to each term in the logistic model are assumed to have chi-squared distribution. An F-test for equality of variances is used to test the hypothesis that the two variances are equal. Table 23 shows that the within sample mean square error for E. washingtonianus and D. excentricus biological toxicity tests were both greater than the treatment mean square error. The graphic demonstrating percent survival for E. washingtonianus (Figure 5) exhibits large variability at Stations 1 and 6, while D. excentricus percent fertilization (Figure 11) was quite variable in sediments from Stations 1, 5, and 6. For this data set, these two tests exhibit more variability in subsamples or laboratory replicates than among true replicates or field replicates. If this is generally the case, then the absence of a “failure” may be due to excessive variability and not due to a true lack of effect. Although not important from a strictly ocean disposal perspective, a large degree of variability may Table 22 Summary of Regression of Tissue Contaminant Levels on Sediment and Porewater Variables Tissue Contaminant Intercept Ag Ammonia Cr Hg Eh Pb PCB pH Salinity Sulphide Total PAHs Zn Ammonia As Cd Cr Cu DryWeight Hg Ni PCB Pb Total PAHs Zn Moisture TOC Percent Sand Percent Silt Percent Clay Goodness of Fit R2 Model P-value Cu Cr 596.1422 -1664.961 -16.164 63.532 0.425 0.573 0.160 0.0814 52 also restrict modeling biological responses as shown in Section 4.2. The coefficients of variation (see Table 24) for reference toxicant tests are examined using Cu and/or Cd, conducted at the Environment Canada Toxicology Laboratories, in Moncton and Vancouver. The reported coefficients of variation are estimated from the LC50s. Pb Ni 3.767 -1217.053 Porewater Variables 0.0321 46.809 Sediment Variables 0.559 0.581 0.0875 0.0781 Zn 286.750 -83.520 0.851 0.00879 PCB Total PAH 116.566 -1382.230 0.687 -0.835 0.687 0.911 0.0415 0.00358 Table 23 Summary of Tests of Equality of Variance Test Species A. virginiana B. proboscidea E. estuarius E. washingtonianus D. excentricus L. pictus P. cornuta R. abronius V. fischeri Table 24 Summary of Reference Toxicant Coefficients of Variation Test Species A. virginiana B. proboscidea E. estuarius E. washingtonianus D. excentricus @10oC D. excentricus @15oC L. pictus P. cornuta R. abronius V. fischeri Exposure temperatures of 15oC are currently being used for the D. excentricus biological toxicity test. Eohaustorius washingtonianus exhibits a large coefficient of variation in reference toxicant tests, but D. excentricus (at 15oC) does not. It would be useful to compare the coefficients of variation from different field studies to see if the findings are validated. 4.4 H4: Does the suite of biological toxicity tests provide a consistent interpretation of the status of the sediment? This hypothesis is tested by testing two subhypotheses. These are: Do all biological toxicity tests characterize the sample in the same way? and Do biological toxicity tests rank the stations in the same way? 53 Degrees of Freedom 14, 84 Test Not Possible 14, 84 14, 84 8, 45 8, 45 Test Not Possible 14, 84 No Laboratory Replication Coefficient of Variation (%) 54.0765 21.687 56.0 40.58 27.57 9.74 45.678 16.293 43.575 27.19 4.4.1 H4a: Do all biological toxicity tests characterize the sample in the same way? This hypothesis may be tested using concordance analysis where the number of “agreements” between pairs of biological toxicity tests is statistically analyzed. A lack of concordance may indicate that constituents of the battery are providing complementary rather than redundant information, which is the raison d’être for a battery, Munawar et al. (1992) and Keddy et al. (1994). 4.4.1.1 Biological Toxicity Tests Pass/Fail Status Relative to Control Sediments (Table 25). The biological toxicity tests are generally in agreement when using the response P-Value 0.9759354 0.5348306 0.02955153 0.04456763 1 0.09214276 Sample Size 33 6 13 3 14 9 13 6 5 138 in control sediments as a point of reference. Stations 1, 5, and 6 show clear evidence of adverse effects while Stations 9, 12, and St. Ann’s Harbour indicate some adverse effects. The most sensitive tests or species in descending order are: D. excentricus, L. pictus = photoluminescent bacteria = A. virginiana = E. washingtonianus, E. estuarius, and R. abronius. The pass/fail data is used to test the null hypothesis H4a: All biological toxicity tests pass or fail stations consistently. Cochran’s test is used (Cochran, 1950) and the assumption is made that the stations were randomly chosen Table 25 Summary of Station Pass/Fail Status* Relative to Control Sediments Test/Species Response Survival Reproduction A. virginiana E. estuarius E. washingtonianus R. abronius D. excentricus L. pictus Luminescence V. fischeri (Solid Phase) Proportion of Tests Failing F 57 * Interim interpretation criteria from Environment Canada (1996) Table 26 Summary of Station Pass/Fail Status* Relative to St. Ann’s Harbour Test/Species A. virginiana D. excentricus E. estuarius E. washingtonianus L. pictus V. fischeri (Solid Phase) R. abronius Proportion of Tests Failing 57 * Interim interpretation criteria from Environment Canada (1996) Stations 1 5 6 9 12 St.Ann’s Proportion of Sites Failing F F F F F F F F F 71 1 5 6 9 12 Proportionof F F F 60 F F F F F 100 F 20 F F F 71 54 40 from among all possible stations. This is true within the criteria for choosing stations (see Materials and Methods, Site Selection Criteria). Cochran’s test statistic of 12.222 on 6 degrees of freedom is associated with a p-value of 0.0572. This suggests that the group of biological toxicity tests passes or fails stations in the same way when the basis for comparison is control sediment. 4.4.1.2 Biological Toxicity Tests Pass/Fail Status Relative to St. Ann’s Harbour (Table 26). The biological toxicity tests show less agreement when using the response in St. Ann’s Harbour 50 16.7 50 0 83.3 F F F 50 F F F F 50 F 71 14 14 29 Stations Sites Failing F F 40 0 60 0 14 14 43 Harbour sediments as a point of reference than when using control sediments. Stations 1, 5, and 6 show evidence of adverse effects while Stations 9 and 12 may indicate adverse effects. The most sensitive tests or species in descending order are: D. excentricus, A. virginiana = photoluminescent bacteria, E. washingtonianus, E. estuarius, and R. abronius = L. pictus. The pass/fail data is used to test the null hypothesis H4a: All biological toxicity tests pass or fail stations consistently. Cochran’s test statistic of 11.636 on 4 degrees of freedom is associated with a p-value of 0.0203. This suggests that the group of biological toxicity tests does not pass or fail stations in the same way. 4.4.1.3 Biological Toxicity Tests Pass/Fail Status Relative to Station 12 Reference Site (Table 27). As before, the biological toxicity tests are generally in agreement and Stations 1, 5, and 6 show clear evidence of adverse effects while Station 9 may be adversely affected. The ordering of the relative sensitivity of species is changed. The most sensitive tests or species in descending order are: E. washingtonianus = A. virginiana = photoluminescent bacteria, L. pictus, R. abronius = E. estuarius and D. excentricus. The pass/fail data is used to test the null hypothesis H4a: All biological toxicity tests pass or fail stations consistently. Cochran’s test statistic of 9.333 on 3 degrees of freedom is Table 27 Summary of Station Pass/Fail Status* Relative to Station 12 Reference Site Test/Species 1 F F A. virginiana D. excentricus E. estuarius E. washingtonianus L. pictus V. fischeri (Solid Phase) R. abronius F F 57 Proportion of Tests Failing * Interim interpretation criteria from Environment Canada (1996) 55 associated with a p-value of 0.0252. This suggests that the group of biological toxicity tests does not pass or fail stations in the same way. As a group, the biological toxicity tests pass or fail stations in the same way only when the basis for comparison is a control sediment. When the pass/fail decision is made relative to a reference sediment, then the group of biological toxicity tests passes/fails sediments differently. This is intuitively satisfying as we know that the response at a control sediment must be “good” or the sediment would not have been chosen as a control sediment. The agreement in pass/fail status when using the group of biological toxicity responses relative to control sediment responses implies that the biological tests are responding adversely to either confounding variables or contaminant effects at the exposure sites. The relationship observed between the amphipods and the PAH gradient (see Section 4.2.3) and other contaminants for those responses which were predictable) suggests that the pass/fail decision is due to the presence of contaminants. The lack of agreement in pass/fail status when using the group of biological toxicity responses relative to reference sediment responses may be due to the effects of factors present at the reference sites. If some organisms exhibit a negative response in the reference sediments, and other organisms do not, the pass/fail status of sediments (relative to the reference stations) will not be homogeneous among the group of toxicity tests. This was the inference or conclusion made Stations 6 5 9 Proportion of Sites Failing F F F F F 75 0 25 75 50 75 25 F F F F 14 43 71 from the three Cochran’s tests on the homogeneity of biological toxicity test responses. Examples of biological toxicity tests performing differently than expected (i.e., nonmonotonically) are seen in the percent survival for E. washingtonianus, photoluminescent bacteria light inhibition, percent fertilization for D. excentricus and L. pictus as shown in Section 3.1. The relative sensitivity of the biological test species changes when reference stations are changed. This observation is on the surface, unsettling. However, when the two species exhibiting reversals in sensitivity, D. excentricus and L. pictus, are examined, it can be seen that the extreme responses for these species were observed in the control sediments, St. Ann’s Harbour or the Station 12 reference site. Thus these stations are acting as pivotal stations, when pass or fail decisions are being made. The choice of station (or pivot) affects the pass/fail decision. 4.4.2 H4b: Do biological toxicity tests rank the stations in the same way? Now that it has been shown that the stations are not passed or failed in the same way using different biological toxicity tests when the basis for comparison is a reference station, it is investigated if the biological toxicity tests at least rank the stations in the same order. The biological toxicity test raw data (i.e., within station responses, not station pass/fail status) from all stations is used, to test the null hypothesis, H4b: Each biological toxicity test ranks the sites equally. The Table 28 Summary of Multiple Comparisons Among Biological Toxicity Tests E. estuarius E. washingtonianus A. virginiana 0.8857, .0553 0.9429, 0.0409 D. excentricus 0.3714, 0.3711 D. excentricus -0.4286, .3067 -0.6571, 0.1252 - E. estuarius - - 0.9429, 0.0409 - - E. washingtonianus L. pictus - - Microtox - - 56 55 Spearman rank correlation coefficient is used to test this hypothesis. Note that normally when many statistical tests are conducted on a data set, the probability of rejecting the null hypothesis when it should not be rejected increases for the group of comparisons as a whole. Often a correction factor is applied to the á value for a single test. For the following data set, the Bonferroni-adjusted á value would be equal to 0.0025. Even when sites are perfectly correlated, the p-value has a theoretical minimum of 0.0298. Thus the null hypothesis can never be rejected when using the Bonferroni adjustment. The pvalues are presented in Table 28 without correction. The table format is “Spearman’s ñ, p-value”. Highlighted cells indicate the pairs of tests that rank the stations in the same way. The D. excentricus, R. abronius, and L. pictus biological toxicity tests, rank stations differently than all the other tests. Only four pairs of the tests rank the stations similarly. The general conclusions reached in this section are: • The biological toxicity tests tend to fail the stations closest to Muggah Creek, irrespective of choice of “reference” sediment. • As a group, the biological toxicity tests pass or fail stations in the same way when the basis for comparison is a control sediment. R. abronius L. pictus Microtox 0.6, 0.2013 -0.08571, 0.7983 1, 0.0298 -0.08571, 0.7983 -0.3714, 0.3711 0.08571, 0.8983 0.4857, 0.3067 -0.2571, 0.5229 0.9429, 0.0409 0.8857, 0.0553 0.8857, 0.0553 -0.3142, 0.4433 - 0.31428, 0.5229 - - 0.6, 0.2013 -0.08571, 0.7983 - - - • The choice of reference station may greatly affect the pass/fail decision rendered by a single biological test. • There is general concordance between tests using pass/fail criteria and the test ranks. Rhepoxynius abronius biological toxicity test is one of the least sensitive tests of those with sufficient data to test, and it tends to rank stations differently from other tests. • The photoluminescent bacterial test ranks stations similarly to E. estuarius and A. virginiana and is similar in sensitivity (as determined by proportion of station “failures”) to A. virginiana and E. washingtonianus. • Amphipods from the genus Eohaustorius rank the stations similarly. 4.5 H5: Do the biological toxicity tests indicate an effect when the TELs or PELs are exceeded? Table 29 summarizes the mean parameter values for those parameters that have interim sediment quality guidelines (ISQGs) established. Values exceeding the probable effect level (PEL) and threshold effect level (TEL) are highlighted, with parameters exceeding the PEL being shaded more darkly. The table shows that all stations would fail, according to the current values. If the mean PEL quotient, given as: i Quotient PEL =.n . . . concentration . PEL i=1 i . n . .. following Long et al., (1998) is used, the probability that stations along the gradient would exhibit toxicity are 56% for Stations 1, 5, and 6, and 24% for Stations 9, 12 and St. Ann’s Harbour. Given all the caveats of the PEL Quotient method (assumes additivity, all contaminants were measured, database derived largely from amphipod survival tests, etc.) Table 57 57 30 shows a good deal of concurrence between expected toxicity using sediment contaminant levels and biological toxicity tests. H5: “Do the biological toxicity tests indicate an effect when the guidelines suggest there should be an effect?” is then tested using concordance analysis to test the agreement between the characterization of a sediment using biological toxicity tests and TELs or PELs. The proportion of parameters exceeding the PEL or TEL is compared with the proportion of stations failing the biological toxicity tests. The data matrix used in estimating the correlations in the first two rows of the summary table is presented in Table 31, as an aid in understanding the rationale behind this analysis. The correlation between the proportions of parameters exceeding the TEL at a given station is estimated (1.00,1.00, 0.95, 0.75, 0.10, and 0.15) with the proportion of biological tests failing a station relative to control sediment (0.71, 0.71, 0.57, 0.29, 0.14, and 0.14). The value of 0.733 in the summary table (Table 31), is the correlation (as measured by Kendall’s Tau) between these two vectors. Statistically, the hypothesis, Ho: “There is no correlation among the proportions using Kendall’s Tau” is tested. Note that this differs from the current paradigm for characterizing a site as failing. Tests significant at the 5% level are highlighted in Table 32. There is a significant correlation between the proportion of failures using biological toxicity tests and TELs and PELs when biological toxicity test inferences are made relative to control sediment. When the biological toxicity test inferences are made relative to the St. Ann’s Harbour station, only the relationship with PELs is significant. The correlation between pass/fail status using biological toxicity tests and TELs is highest when biological toxicity tests inferences are made relative to the Station 12 reference site. This correlation is not significant due to the low number of data points (four pairs of pass/fail decisions) available for the comparison. As* Cd* Cr* Cu* Hg* Pb* Zn* Table 29 Summary of Sediment Contaminants with PELs and TELs 1 Parameter Metals (g/g dry weight) 41.0000 1.16667 81.2333 101.3333 0.7077 285.6667 516.2667 PCBs (ng/g dry weight) 2095.0230 total PCB* PAHs (ng/g dry weight)* 1161.3778 2-Methyl- Naphthalene Acenaphthylene 690.1267 419.2858 Acenaphthene 1636.1570 Fluorene 8839.2643 Phenanthrene 5498.4707 Anthracene 13651.301 19589.421 14229.563 Fluoranthene Pyrene Benzo(a) Anthracene Chrysene 17391.319 23464.113 Benzo(a)Pyrene 4223.4349 14.85907 Dibenzo(a,h) Anthracene Mean PEL Quotient** * CCME (1999); ** Long et al. (1998). 3 Negative numbers reflect values below the detection limit. The detection limit is a function of the instrument response, the noise in the chromatogram, and the amount of sample used. 5 39.3333 0.9267 86.5667 73.6667 0.4863 214.0000 865.6667 1186.2303 710.9085 372.3373 259.1971 901.9746 4929.8272 2787.5911 6054.4574 6888.9134 6848.0221 8184.1967 8502.3410 1871.4732 7.593556 Stations 9 6 10.0000 32.6667 0.1467 0.4667 33.9000 61.7000 22.7000 53.7000 0.0377 0.3330 32.0000 133.3333 91.1000 281.5667 642.7346 -69.13103 88.1666 361.3295 7.2086 202.4355 27.6189 152.4286 50.3778 432.1399 340.4782 2427.1508 125.8528 1362.9712 337.5117 289.4181 283.8230 2950.9182 3450.7664 3147.7763 312.2847 3798.1966 193.5271 3352.9856 44.7249 671.2666 0.438296 3.619002 12 St. Ann’s Harbour 15.6667 9.6667 0.2467 0.08000 41.9333 24.1000 37.3000 13.0000 0.0550 0.0243 37.0000 21.0000 84.2333 56.2000 -63.5605 -30.8474 5.5598 39.7952 0.0000 1.1304 0.8682 2.6713 5.1443 12.4128 13.3471 50.5420 9.3728 41.7342 31.9409 24.5577 22.4324 41.2303 34.9358 24.9308 23.0015 30.7719 17.8804 11.3748 0.0000 0.0000 0.108411 0.107492 58 58 PEL TEL µg/g dry µg/g dry 41.6 7.24 4.2 0.7 160 52.3 108 18.7 0.70 0.17 112 30.2 271 124 ng/g ng/g dry 189 21.5 ng/g dry ng/g dry 201 20.2 128 5.87 88.9 6.71 144 21.2 86.7 54.1 245 46.9 1494 1398 693 113 153 74.8 846 108 763 88.8 135 6.22 Table 30 Comparison of Mean Quotient PELs and Observed Toxicity Probability that sediment is acutely toxic (Long et al., 1998) Proportion of Toxicity tests Failing relative to control sediments (from Table 25) Table 31 Sample Data Set Assessment Tool Proportion Samples > PEL (from Table 29) Proportion Samples > TEL (from Table 29) Proportion Stations “Failing” relative to Control Sediment (from Table 29) Table 32 Summary of Tests of Correlation Between Proportions of Stations Failing Assessment Tool Comparison with: Biological Toxicity Tests Relative to: Control Sediment TEL PEL St. Ann’s Harbour TEL PEL Station 12 Reference Site TEL PEL Additionally, the proportion of tests failing the sediments, concurs with the probability that a sediment is highly toxic (see Long et al., 1998 for definition), based upon mean PEL quotients. 4.6 H6: Do the biological toxicity tests indicate an effect when the in-situ benthic macroinvertebrate community does? This hypothesis is tested by examining the data set for patterns. Ordinations of the benthic macroinvertebrate abundances are used to explore structure in the data set (Figure 27). Ordination of averaged raw abundances using the correlation matrix showed that a large proportion of the variability in the data set was due to simple numerical dominance in the case of the first 59 59 Station 1 5 6 9 12 St.Ann’s Harbour 24 24 24 56 56 56 14 14 29 57 71 71 Station 1 5 6 9 12 St.Ann’s 0.80 0.75 0.75 0.05 0.00 1.00 1.00 0.95 0.75 0.10 0.71 0.71 0.57 0.29 0.14 Harbour 0.00 0.15 0.14 P-value Kendall’s Tau 0.0293 0.0266 0.1184 0.0374 0.071 0.279 0.733 0.733 0.6 0.8 0.833 0.5 principal component and the presence of specific organisms for the second principal component. The data is well summarized by two components accounting for 82.5% of the total variability in the data set. Plots of station scores show that the first principal component strongly separates Stations 9 and 12 from the other stations. The second principal component strongly separates out the Station 12 reference site and only weakly separates Station 9. Both principal components group Stations 1, 5, 6 and St. Ann’s Harbour together. Also, Arenicola Marine (1997) describes the St. Ann’s Harbour reference station as being most similar with respect to benthic community structure, to Station 1 (Arenicola Marine, 1997). St. Ann’s Harbour was also characterized as one of the two most affected sites, based on benthic community structure ordination which groups Stations 1, 5, 6 and St. Ann’s Harbour together, based on raw abundances. The result of the difference is likely the 4th root-transformation that greatly reduces the effects of numerical abundance. The interpretation of the ANOSIM procedure following the ordination is that once the effects of numerical abundance have been reduced, all stations differ from either of the reference stations, likely on the basis of species composition. Undue weight should not be placed upon this finding until the literature corroborates the use of this test with other procedures and data sets. If the number of times that both biological toxicity tests fail a station (based upon the failure of one acute amphipod test) and benthic communities fail a station is counted, Table 33 can be created. Fisher’s exact test is used to test the null hypothesis that the two row and column variables are independent. The p-value for the both tables is 1. Consequently, it is found that the benthic macroinvertebrate community and the biological toxicity tests do not characterize the sediments in the same way. 4.7 H7: Do the three evaluation tools characterize sediments in the same way? The degree of agreement in the classification of a sediment or site using the three characterization Table 33 Concordance Between Station Characterizations using Benthic Community Structure and Biological Toxicity Tests Biological Toxicity Tests (relative to St. Ann’s Harbour) Fail Pass Biological Toxicity Tests (relative to Station 12) Fail Pass 61 61 tools is summarized in Table 34. The stations fail a biological toxicity test if more than 1 test fails or a single amphipod toxicity test fails. There is perfect agreement between the biological toxicity tests and the sediment PELs, based upon pass/fail status. (Note that the current use of TELs in the ocean disposal context is to trigger a tier 2 assessment, not necessarily prevent dredged material from being disposed of at sea.) The benthic macroinvertebrate community and the TELs concur. However, the ANOSIM procedure used to compare the benthic community macroinvertebrate structure between stations does not agree with the ordination of the benthic macroinvertebrate community structure likely due to the difference in emphasis on numeric abundance. We refrain from making conclusions regarding the concordance of the three assessment tools pending a comparison of the three data sets using the raw data. (See the following final hypothesis.) 4.8 H8: How strongly are the three data sets correlated? The study design is a gradient design using a sediment quality triad approach. The constituents of the triad are biological toxicity tests, sediment physical/chemistry and in-situ benthic macroinvertebrate community structure. The degree of correlation between these data sets is explored using ordination (principal components analysis) and permutation tests (Mantel’s test). Benthic Macroinvertebrate Community Pass Fail 0 0 3 2 Benthic Macroinvertebrate Community Pass Fail 0 0 3 1 Table 34 Summary of Station Pass/Fail Status* Using All Criteria Biological Toxicity Tests Relative to Control Sediment Biological Toxicity Tests Relative to St. Ann’s Harbour Biological Toxicity Tests Relative to Station 12 Benthic Macroinvertebrate Community F F Criteria for biological toxicity test given in Environment Canada (1996). Sediment TEL Sediment PEL * ** By definition, the reference site cannot “fail”; therefore, the pass/fail status of the reference station (either St. Ann’s Harbour or station 12) is not in question. *** Fails when compared to St. Ann’s Harbour. We begin with ordinations of the sediment physical/chemistry and biological toxicity test responses. It was previously shown that individual PAHs are highly correlated with total PAHs, consequently total sediment and porewater PAHs will be used in lieu of individual sediment and porewater PAHs, respectively, for the ordinations (see Figure 28). The principal components analysis shows that the sediment physical/chemical data set is highly structured. Three principal components describe 80.200% of the total variability in the data set. The first principal component orders the stations in the following order: 1, 5, 6, St. Ann’s Harbour, 9 and 12. There is clear evidence of the gradient by the loadings (not shown) on the following 12 variables in order of magnitude: TOC, sediment Cu, Pb, As, Cr, porewater PCBs and PAHs, sediment PAHs, sediment Hg, Ni, total PAHs, and Cd. The St. Ann’s Harbour station has intermediate values for these parameters; hence it’s position in the ordering due to the first principal component. The other stations are arranged in an order corresponding to the known gradient in PAHs and metals. The second principal component describing an additional 18.6% of the variability in the data set separates St. Ann’s Harbour from the other stations. This is due to loadings (not shown) that contrast porewater Zn, Cr, sulphide, sediment ammonia, moisture and redox potential with porewater pH and ammonia, dry weight, 62 62 Pass/Fail Status 1 5 6 9 12 St. Anne’s Harbour P NA** P NA** P P NA** F***/ P P P F F F F F F F F F F F F F NA** F P F P F P F F F F porewater Ag, salinity, and Hg. The separation of the St. Ann’s Harbour station from the other stations is due to the low values of pH, porewater ammonia, and dry weight (relative to the other stations). The low pH may have increased the porewater levels of Zn and Cr (Sigg, 1987). Porewater levels of these two metals are the highest in St. Ann’s Harbour stations, even though these sediments contain only moderate amounts of sediment-associated Zn and Cr (relative to the other stations). The third principal component describing an additional 8.01% of the variability in the data set separates Station 9 (not shown) from the other stations due to contrasting loadings on fines (clay and silt) and sand. Station 9 has the largest amounts of clay and silt and the smallest amount of sand relative to the other stations. Thus we see that the stations are ordered along a concentration gradient, St. Ann’s Harbour sediments are moister, (possibly with increased porosity) and have elevated porewater metal levels, possibly due to pH. Station 9 is anomalous with respect to sediment grain size composition containing more clay and less sand than the other stations. The biological toxicity test responses are ordinated using principal components analysis (PCA). Three components describe 94.600% of the total variability in the data set. The scores plot is shown in Figure 29. Se co nd Pr -2.00 -4.00 -6.00 inc ip al Co m po ne t Figure 28 -4.00 2nd Principal Componen Scores Plot for Sediment Physical/Chemistry Variables 12 St. Ann's Harbour 6 5 9 1 -2.00 -3.00 4.00 3.00 2.00 1.00 0.00 0.00 -1.00 -1.00 -2.00 -3.00 1st Principal Component 1.00 Scores Plot for Biological Test Responses Figure 29 63 63 3.00 2.00 1.00 2.00 0.00 0.00 -1.00 -2.00 -3.00 -4.00 -5.00 -6.00 First Principal Component 8.00 6.00 4.00 1 5 6 9 12 St. Ann's Harbour 3.00 2.00 Table 35 Summary of Ordinations Ordination on: Benthic Community Structure Biological Toxicity Tests Sediment Physical Chemistry First component separates Stations 9 and 12 due to numerical abundance of specific taxa. The second components separates the Station 12 reference site due to the presence of organisms found only at that station. Stations 1, 5, 6, and the St. Ann’s Harbour station are grouped together. Shows gradient, separates St. Ann’s Harbour and weakly separates Station 9 due to contrasts in biological toxicity test responses. Stations 1, 5, 6, and 12 are grouped together. Shows gradient, separates St. Ann’s Harbour due to porewater ammonia, metals and low pH, and separates Station 9 due to sediment grain size differences. Stations 1, 5, 6, and 12 are grouped together. Table 36 Summary of Mantel’s Test Comparisons Comparison Biological Toxicity Tests and Benthic Community Structure Sediment Physical/Chemistry and Benthic Community Structure Sediment Physical/Chemistry and Biological Toxicity Tests Also, the proportion of tests failing the sediments, concurs with the probability that a sediment is highly toxic (see Long et al., 1998 for definition), based upon mean PEL quotients as shown in Table 30. 4.9 H9: Do the interim biological toxicity test interpretation criteria need to be adjusted to match assessments of sediment quality using benthic community structure or SQGs? The previous hypothesis tests indicate that the interim biological toxicity test interpretation criteria characterize (i.e., pass/fail) sediments in the same way as the PELs. When the biological toxicity test battery is used to determine the suitability of dredged materials for ocean disposal, the battery will consist of one each, of the following tests: amphipod survival, echinoderm fertilization, bivalve bioaccumulation, and the Microtox® (V. fischeri) solid-phase test, (L. Porebski, pers. comm., Environment Canada, Ottawa, ON, 1999). A sediment will fail the biological test battery if two or more tests fail or if one acute toxicity test (amphipod) fails. 65 65 Comment P-Value 0.364 0.892 0.199 Correlation 0.133 -0.205 0.436 The significant difference, or in the terms of this report, the “failure” of Stations 1, 5, 6 and 9 relative to either of the reference stations based upon the in-situ benthic macroinvertebrate communities should not be given undue weight. The ANOSIM procedure is relatively new and may be overly sensitive. Ordination procedures and expert opinion group Stations 1, 5, 6 and St. Ann’s Harbour together. This grouping of the reference station with the most highly contaminated sites is contrary to the findings of the biological toxicity tests and sediment contaminant concentrations. The proportion of biological toxicity tests failing a station is more highly correlated with the proportion of stations failing according to PEL criteria than TEL criteria. Also, the pass/fail status of stations is identical using biological toxicity tests and sediment PELs. Finally, the proportion of tests failing the sediments concurs with the probability that a sediment is highly toxic based upon mean PEL quotients. Empirically and statistically that is seen as a group, the interim biological toxicity test criteria reflect sediment contamination in a meaningful way. Section 5 Summary of Hypothesis Tests At times during the report, a hypothesis test or analysis is repeated when one of the two reference stations is used as a basis for comparison. If changes in inferences resulting from the use of a different reference station occur, the differences are summarized in Table Table 37 Summary of Hypothesis Tests Hypothesis Test H1: Homogeneity of confounding factors across station. H2a: Effect of confounding factors on biological toxicity test responses. H2b: Are the dose responses predictable? H3: Do biological toxicity tests perform consistently within a site? Median TOC, percent moisture, sand, silt clay, redox potential, porewater and sediment ammonia and sediment sulphide values are different for at least one of the stations along the gradient. Of the organisms that reacted with a strong biological response across the gradient none were best predicted by confounding factors. However, TOC was often highly correlated with those variables included in the model and was therefore a contending explanatory variable. A strong response to the PAH gradients occurred among the amphipods A. virginiana and E. estuarius. Other organisms responded to different contaminants such as metals. The exploration of the various models not presented, shows that non-contaminant variables are also associated with the responses, but not as strongly as with contaminants. The correlation between contaminants and non-contaminant factors, particularly TOC and redox potential was often high. Non-contaminant effects such as grain size were not observed. The within sample mean square error or laboratory replicate variability for E. washingtonianus and D. excentricus biological toxicity tests were both greater than the treatment mean square error or field replicate variability. 66 67 37 by the inclusion of a separate set of rows for the hypothesis so affected. The results and implications of hypothesis 2 reflect the manual or consensus model building approach rather than the automated model building approach of Section 4.2.2. Result Implication Biological toxicity test responses and benthic community structure may vary due to confounding effects as well as toxicant effects. Confounding factors do not affect the observed responses more than the contaminant variables. Confounding factors, particularly TOC are often associated with high levels of contaminants. Confounding factors should continue to be monitored. The biological toxicity tests are responding primarily to the contaminants. These tests exhibit greater variability within samples than among samples. Hypothesis Test H4: Does the suite of biological toxicity tests provide a consistent interpretation of the status of the sediment? H4a: All biological toxicity Using pass/fail criteria, L. pictus and R. tests pass or fail stations abronius were the least sensitive species, D. consistently. (St. Ann’s excentricus was the most sensitive. Harbour as reference Statistically, the group of biological toxicity station). does not pass or fail stations in the same way. H4a: All biological toxicity Using pass/fail criteria, D. excentricus and R. tests pass or fail stations consistently. (Station 12 as reference station). abronius were the least sensitive test species while A. virginiana, E. washingtonianus and V. fischeri were the most sensitive organisms. Statistically, the group of biological toxicity tests does not pass or fail stations in the same way. H4a: All biological toxicity Using pass/fail criteria, R. abronius was the tests pass or fail stations least sensitive species while D. excentricus consistently. (Control was the most sensitive biological toxicity test. sediment as reference Statistically, the group of biological toxicity station). tests passes or fails stations in the same way. H4b: Each biological R. abronius, L. pictus and D. excentricus rank toxicity test ranks the sites the stations differently from the other equally. biological toxicity tests. H5: Do the biological toxicity tests indicate an effect when the TELs or PELs are exceeded? H6: Do the biological toxicity tests indicate an effect when the in-situ benthic macroinvertebrate community does? 67 67 Result See following subsections. A comparison of the proportion of biological toxicity tests eliciting a fail response at a station with the proportion of parameters exceeding the PEL or TEL was significant. The degree of correlation between station pass/fail status using biological toxicity tests and chemical values was higher using PEL than TEL. The biological toxicity tests always indicate an effect (based upon the failure of at least one test. The benthic macroinvertebrate community stations always differ from either of the reference stations using permutation methods but not when using ordination methods. Implication The basis for comparison affects the pass/fail decision and consequently the rankings of relative sensitivity. The biological toxicity tests as a group, only pass or fail stations in the same way when the basis for comparison is a control sediment. The choice of the basis for comparison is important in an ocean disposal context. The choice of biological toxicity test (and species) influences the pass/fail status of a sediment/station. The choice of biological toxicity test influences the pass/fail status of a sediment/station. The choice of biological toxicity test does not influence the pass/fail status of a sediment/station. Three biological toxicity tests ranked the stations differently. The most and least sensitive species ranked the stations differently than the other biological toxicity tests. Biological toxicity test responses are triggered when TEL or PEL values are exceeded. The basis for comparison affects the pass/fail decision and consequently the strength of the relationship between PELs or TELs. The proportion of biological toxicity tests responding, roughly concurs to the proportion expected to respond based upon mean PEL quotients. Since all stations “failed” according to both sets of criteria this observation is equivocal. The interpretation of the ANOSIM procedure following the ordination is that once the effects of numerical abundance have been reduced, all stations differ from either of the reference stations, likely on the basis of species composition. Hypothesis Test H7: Do the three evaluation tools characterize sediments in the same way? H8: How strongly are the three data sets correlated? H9: Do the interim biological toxicity test interpretation criteria need to be adjusted to match assessments of sediment quality using benthic community structure or SQGs or ISQGs? 68 68 Result The three evaluation tools do not characterize the sediments in the same way. The benthic community fails all stations (see caveats in Section 4.6) as would TELs if employed as pass/fail criteria. There is perfect agreement between pass/fail status using biological toxicity tests and PELs (if PELs are employed as pass/fail criteria). Ordinations using biological toxicity test data and sediment physical chemistry, order the exposure stations along the chemical gradient and separate both Stations 9 and St. Ann’s Harbour. The ordination on benthic community structure separates Stations 9 and 12 while grouping Stations 1, 5, 6 and St. Ann’s Harbour together. Mantel’s tests show no significant correlation between the three data sets using Euclidean distances. The interim biological toxicity test interpretation criteria reflect a probable effect level. Implication Sediment chemistry and biological toxicity tests characterize sediments in the same way. The ordinations on biological toxicity test data and sediment physical chemistry indicate that the biological toxicity tests reflect the sediment chemistry even though this is not corroborated by Mantel’s test. The ordination on the in-situ benthic community do not order the stations along the known gradient but does group Stations 1, 5, and 6 together. However, the reference station is also grouped with these contaminated sites. The current interpretation criteria afford a short-term or lethal level of protection to the environment based upon the responses measured, for the types and concentrations of contaminants encountered. Note: This study used relatively short-term exposures and a small group of taxa. Further investigation would be advisable to clarify the links with ecosystem–level bioaccumulative or sublethal effects. Section 6 Discussion 6.1 Performance Evaluation of Biological Toxicity Tests 6.1.1 Acute Survival Tests. The acute survival tests using amphipods show a dose response ranging from moderate to very strong. The responses of A. virginiana and E. estuarius are related to sediment and porewater PAHs, respectively. The pass/fail comparisons with the control sediments show that the amphipods as a group are less sensitive than echinoderms and V. fischeri. When the pass/fail characterizations are made relative to the St. Ann’s Harbour reference station, or the Station 12 reference site, the amphipods are of intermediate sensitivity, relative to echinoderms and V. fischeri. Stations are not passed/failed in the same way using the different biological toxicity tests. Note that this evaluation does not include polychaetes, due to the lack of replication. This conclusion holds whether pass/fail decisions are made relative to St Ann’s Harbour or the Station 12 reference site. When the biological test responses are used rather than pass/fail status, it is found that A. virginiana, E. estuarius, and E. washingtonianus rank the stations similarly. Of the amphipod species tested, the use of A. virginiana and E. estuarius is recommended based on the general concurrence of pass/fail status, stability of sensitivity to changing reference conditions, and statistically acceptable variability and response to the known organic contaminant gradient in this study. 6.1.2 Sublethal Tests. No dose response was predictable for D. excentricus, L. pictus, B. proboscidea, or P. cornuta. The logtransformed dose response for V. fischeri was well predicted by sediment-associated Ni. When the pass/fail status of sediments relative to control sediments is determined using sublethal tests the echinoderms are the most sensitive 69 69 biological toxicity tests, with the V. fischeri test being of intermediate sensitivity. When the pass/fail test is made relative to the St. Ann’s Harbour sediments, the D. excentricus fertilization assay remains among the most sensitive tests, while the L. pictus biological toxicity test becomes the least sensitive, and relative sensitivity of the V. fischeri test remains unchanged. When the pass/fail test is made relative to the Station 12 reference site sediments, the D. excentricus fertilization assay is grouped among the least sensitive tests while the other two sublethal tests are of intermediate sensitivity. Among the sublethal tests, the within sample variability for D. excentricus was greater than the variability among samples. This implies that differences between stations or between exposure and control sites may be obscured by the variability within a site. Of the sublethal tests, the echinoderm tests are the most problematic in this study. Large statistical variability was shown by the D. excentricus fertilization assay. The echinoderm pass/fail tests were also sensitive to the choice of reference station. Porebski et al. (1998) also found the D. excentricus test to be variable and sensitive to the choice of reference station. The V. fischeri pass/fail test remained consistent when reference stations were changed, as the pass/fail decision is not made relative to performance at a reference site. Section 6.6 includes suggestions for modifying interpretation criteria for this test that would affect its consistency. Polychaete survival was not a function of the gradient and only P. cornuta growth exhibited any systematic change along the gradient. However, this change in P. cornuta growth, was not predictable. Also, P. cornuta did not respond as anticipated to a known metal gradient (Porebski et al., 1998). Problems with the sublethal toxicity tests based on the results of this study are the general lack of predictable dose responses along a known contaminant gradient (with the exception of the V. fischeri toxicity test) and the equivocal pass/fail status of sediments based upon echinoderm toxicity tests. 6.1.3 Bioaccumulation Tests. The bioaccumulation of porewater PAHs and porewater PCBs was highly predictable in M. nasuta. The bioaccumulation test using M. nasuta is being standardized by the USEPA and is advocated as a bioaccumulation test for the evaluation of dredged materials for disposal at sea (USEPA/USACE, 1998). The Canadian species M. balthica was used to evaluate bioaccumulation along a known metals gradient (Porebski et al., 1998) but the laboratories involved commented that the test was labour intensive due the small size of the organism. M. nasuta may provide a better alternative in light of it’s larger size, demonstrated performance under an organic contaminant gradient and use by the USEPA in sediment quality assessment which allows a broader basis for comparison. However, M. nasuta is not native to the colder waters of Canada. Before a final decision is made to endorse M. nasuta over M. balthica, comparative sensitivity and variability studies should be undertaken. Sediment Physical Chemistry 6.2 6.2.1 PELs/TELs and Biological Toxicity Tests. Contaminant levels alone do not fail a station in the ocean disposal context. However, a correlation between guideline levels and biological toxicity test failures provides some degree of assurance that biological effects criteria do not require adjustment. In this study there was a significant correlation between the proportion of failures at a given station using biological toxicity tests and both TELs and PELs when biological toxicity test pass/fail decisions are made relative to control sediment, but only with PELs when St. Ann’s Harbour was the 70 basis for biological toxicity test pass/fail decisions. The proportion of tests failing the sediments (pass/fail decision made relative to control sediment) concurs with the probability that a sediment is highly toxic based upon mean PEL quotients. 6.2.2 Porewater and Sediment Chemistry. The issue of using porewater chemistry in addition to sediment chemistry has been raised. In this study, the bioaccumulation of porewater PAHs and porewater PCBs was more predictable in M. nasuta than sediment concentrations of PAHs and PCBs. Of the two strong dose responses observed over the organic contaminant gradient, one was well predicted by sedimentassociated PAHs (A. virginiana survival) while the other was well predicted by porewaterassociated PAHs (E. estuarius survival). The porewater biological fertilization tests using echinoderms did not respond in a predictable manner to either the porewater or sediment measured contaminants. Automated model building methods suggest that porewater variables explain more of the observed variability in the biological toxicity test responses than sediment variables. However, this observation was not confirmed when modelling dose-responses using a hands-on approach. The work conducted during the course of this study indicates that porewater variables are correlated with some biological test responses. An assessment of the additional benefits that the use of porewater contaminant measurements would bring to the ocean disposal program is beyond the scope of this project. However, it should be noted that other jurisdictions acknowledge the relevance of porewater contaminant concentrations in assessing sediment. The USEPA is developing sediment quality guidelines using an equilibrium partitioning approach (EqP) (USEPA 1992; 1999) for nonionic organic compounds and simultaneously extracted metals (SEM) for metals criteria (Ankley et al., 1996). Both of these methods are porewater methods in that they address porewater contaminant concentrations indirectly (EqP) or directly (SEM). The inclusion of porewater variables (in some form) would allow for comparison with other sediment assessment methods and may be of use in understanding the results of porewater tests. 6.2.3 Total PAHs versus Individual PAH Measurements. Individual PAHs were highly correlated with total PAHs. This is in part due to the fact that each PAH contributes to the measurement of PAH. However this observation has also been made in investigations of freshwater sediments contaminated with organic compounds (Moran et al., 1997). Swartz (1999) and USEPA (1999) discuss the utility of total PAH measurements as a surrogate for individual PAH measurements. Based on this and other recent studies, the disposal at sea program may wish to continue to rely on a total PAH value for screening sediments. 6.3 Choice of Reference Stations As a group, the biological toxicity tests do not pass or fail stations in the same way when the basis for comparison is a reference sediment. When the pass/fail decision is made relative to a control sediment, then the group of biological toxicity tests passes/fails sediments in the same way. The lack of agreement in pass/fail status when using the group of biological toxicity test responses relative to reference sediment responses may be due to the effects of factors present at the reference sites. If some organisms exhibit a negative response in the reference sediments, and other organisms do not, the pass/fail status of sediments (relative to the reference stations) will not be homogeneous among the group of toxicity tests. The practical result in an ocean disposal context may be contradictions within the test battery. Another consequence of organisms responding differently to reference sediments is the observed change in relative sensitivity of the biological test 71 species. This observation is on the surface, unsettling. However, when the two species exhibiting reversals in sensitivity, D. excentricus and L. pictus, were examined, the extreme responses for these species were observed in the control sediments, St. Ann’s Harbour or the Station 12 reference site. Thus these stations are acting as pivotal stations, when pass or fail decisions are being made. The choice of station (or pivot) affects the pass/fail decision. The agreement in pass/fail status when using the group of biological toxicity responses relative to control sediment responses implies that the biological tests are responding adversely to either confounding variables or contaminant effects at the exposure sites. The choice of reference station also affected the strength of the correlation between the proportion of sediment “failures” using biological toxicity tests and proportion of sediment “failures” using TELs or PELs. The strongest correlation between biological toxicity test pass/fail decisions and sediment chemistry pass/fail decisions generally occurred when the control sediment was used as a reference station. There may be some circularity in this observation as PELs and TELs are derived using the geometric mean from percentiles of a biological toxicity test effects and no effects database (EC, 1995b). The circularity of the observation arises as the definition of “effect” and “no effect” may be made relative to a reference or control sediment specific to each experiment. The determination of effect or no effect for each of the individual data sets comprising the effects/no effects database was made by the contributing author, and may have been made relative to a reference or control sediment. A reference station should be chosen in the context of the experimental/study goals. For this study, the reference station was chosen to match known confounding factors, such as sediment grain sizes and station depths in Sydney harbour in order to explore issues arising when interpreting the pass/fail status of sediments using three assessment tools. Although most potential confounding factors did vary statistically among stations, the confounding factors were not primarily responsible for predictable biological toxicity test responses. Thus, the observed differences in mean values of confounding factors were statistically significant, but not ecologically significant. This observation implies that confining confounding factors to class intervals or ranges, as is done for grain size criteria in amphipod biological toxicity tests (see Table 6) is sufficient to ensure comparability of biological toxicity test results between exposure and reference stations. Class intervals or ranges for a given confounding factor within which a biological test would not significantly vary could be used to better select reference sites. The key confounding factors for which class size restrictions should be developed will likely be species-specific. When choosing a reference site for an ocean disposal permit application, the USEPA (1991) suggests that the grain sizes at the reference station(s) be as similar as is practical to the grain sizes of the dredged material, and that the reference station reflects conditions prevailing at the disposal site, before disposal. This concurs in principle with the suggestion to only compare sediments when known confounding factors fall within the same class interval. Although the reference site in this study was chosen with due diligence, it did not reflect the reference condition for the in-situ benthic macroinvertebrate community but did for the biological toxicity tests being evaluated (with the exception of L. pictus). Given that the goals of the study were to explore issues arising when interpreting the pass/fail status of sediments using three assessment tools, the choice of St. Ann’s Harbour as a reference station should stand. However, the discrepancy between responses observed at different levels of biological organization (i.e., single organism toxicity tests and benthic macroinvertebrate community structure) suggests that critical elements in reference site selection were not identified in advance. 72 The two reference stations were chosen based on similarity of known confounding factors with exposure sites (see Section 2.1). Upon chemical analysis, the St. Ann’s harbour station exceeded Cu and Pb TELs while the Station 12 reference site exceeded TELs for 2-Methyl-Naphthalene and As. Similarly, Porebski et al. (1998) found that the reference site for Belledune harbour exceeded at least one TEL. These two observations highlight the practical difficulties in choosing a reference station. It is unlikely that many reference stations could be found that did not exceed at least one TEL. The importance of reference site selection should not be under-emphasized, especially in the case of ocean disposal permit application, as this is essentially an experiment with two “treatments”, namely a reference sediment and a sediment being considered for ocean disposal. Good experimental design suggests equal experimental effort be expended upon both “treatments” in the absence of a priori knowledge. Thus an equal number of reference and exposure sites represents the best allocation of sampling effort. Note that multiple samples from one reference site or area, do not constitute replicates. These samples are best described as pseudoreplicates (Hurlbert, 1984), and have been decried as not being representative of natural variability (Underwood, 1991). The use of multiple reference stations to determine potential contaminant effects is endorsed by the USEPA (1991) in an ocean disposal context and Environment Canada (1998b) in the context of assessing potential pulp and paper mill environmental effects. Some problems do exist if multiple reference sites are used as a basis for comparing biological toxicity test results. These include restrictions on choice of reference site to avoid biasing a permitting application, the availability of one, let alone several suitable reference sites, the presence of unsuspected contaminants and the additional costs incurred. 6.4 Reference Stations versus Control Stations The basis of comparison affected: 1) the pass/fail decision for a sediment; 2) the consistency of interpretation of the biological toxicity tests; and 3) the degree of correlation between biological toxicity test pass/fail assessments and sediment contaminants. 1) The choice of reference station affected the pass/fail decision rendered by a single biological test. 2) As a group, the biological toxicity tests pass or fail stations in the same way when the basis for comparison is a control sediment but not when the basis for comparison are reference sediments. 3) The biological pass/fail assessments concur with those made using sediment chemistry values when the biological pass/fail assessments are made relative to control sediment. When reference sediment is used as the basis for a making pass/fail decision with a biological test, the biological test pass/fail assessments are less strongly correlated with the sediment assessments made using contaminant criteria. This occurs as the proportion of biological tests failing the stations along the gradient does not decrease monotonically when the basis for biological pass/fail decisions are the reference stations. The number of failures triggered by TEL/PEL values does decrease monotonically along the gradient. 6.5 Comparison of Three Sediment Characterization Methods Ordinations show that all three assessment tools group Stations 1, 5, and 6 together. Station 9 is separated from the other stations on the basis of numerical abundance of specific taxa, contrasts in biological toxicity test responses, and sediment grain size differences. The ordinations using biological toxicity test responses and sediment 73 physical chemistry also separate St. Ann’s Harbour from other stations. This finding concurs with the observed correlation between the pass/fail characterizations using biological toxicity tests and TELs or PELs and the observed bioaccumulation of organic compounds in bivalves. The lack of correlation between the health of the in-situ benthic macroinvertebrate community and sediment contamination may be more a consequence of an effect at the St. Ann’s Harbour reference station than a lack of effect, lower in the gradient. Figure 18 shows that the benthic macroinvertebrate community richness increases almost monotonically until the St. Ann’s Harbour reference station is reached. The failure of the benthic community at this single station is enough to mask this trend when comparing ordinations and may in part be responsible for the lack of significance when using Mantel’s test. 6.5.1 Interpretation Criteria. Sediment contaminant TELs and PELs do concur with biological toxicity test results, with PELs being more highly correlated with biological toxicity test responses than TELs (note that in an ocean disposal context a station does not fail on the basis of sediment chemistry alone). Empirically and statistically it can be seen that as a group, the current interim biological toxicity test criteria reflect sediment contamination in a meaningful way. The current interpretation criterion for the bivalve bioaccumulation test is a statistically significant difference between reference or control sites and exposure sites. The pass/fail status of the test was determined, but should be interpreted cautiously due to a lack of replication. The test did respond well to the organic contaminant gradient. Models exist demonstrating the link between sediment organic contaminants (PCBs and DDTs) and bioaccumulation in M. nasuta (Boese et al., 1997). Guidelines based upon risks to human health are available for some contaminants (USEPA/USACE, 1998). Health Canada or References Allen, H.E., G. Fu, and, B. Deng. 1993. Analysis of acid-volatile sulfide (AVS) and simultaneously extracted metals (SEM) for the estimation of potential toxicity in aquatic sediments. Env. Tox. Chem., 12:1441–1453. Ankley, G. T., D.M. Di Toro, D. J. Hansen, and W. J. Berry. 1996. Technical basis and proposal for deriving sediment quality criteria for metals. Env. Tox. Chem., 15:2056–2066. Arenicola Marine. 1997. Sample collection and analysis for benthic community structure: Sydney Harbour, NS, July 1997. Atkinson, A. C. Plots, transformations, and regressions. Oxford Science Publications, Oxford, England, 282 pp. BEAK Consultants Ltd. 1996. Development of interpretive guidance for biological tests using chemical gradient studies. Phase 1: Confirmation study. report prepared for Environment Canada, Marine Environment Division, Ottawa, ON. Boese, Bruce L., Henry Lee II, and S. Echols. 1997. Evaluation of a first-order model for the prediction of the bioaccumulation of PVBs and DDTs from sediment into the marine depositfeeding clam, Macoma nasuta. Environmental Toxicology and Chemistry, 16:1545–1552. British Columbia Ministry of Environment, Lands and Parks. 1994. Sand dollar sperm inhibition bioassay manual. Unpublished report. CCME (Canadian Council of Ministers of the Environment) 1999. Canadian Sediment Quality Guidelines. CCME, Winnipeg, Manitoba. (CEPA) Canadian Environmental Protection Act. 1999. Statutes of Canada, Chapter 33. Federal Government of Canada, Ottawa, ON, Canada. 76 Chapman, P.M. 1992. Sediment Quality Triad Approach. In: USEPA, Sediment Classification methods compendium. 10-1-10- 18, Office of Science and Technology, Washington, DC. Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol., 18:117–143. Cochran, W.G. 1950. The comparison of percentages in matched samples. Biometrika, 37:256–266. Collett, D. 1991. Modelling binary data. Chapman and Hall, Toronto, Ontario, 369 pp. Devitt, S. 1997. Field activity report for collection of marine sediments from Sydney Harbour. Prepared for Ms. L. Porebski, Environment Canada, Marine Environment Division, Ottawa, Ontario. Edwards, D. and J. J. Berry. 1987. The efficiency of simulation-based multiple comparisons. Biometrics, 43:913–928. EC (Environment Canada). 1992. Analytical methods for organochlorinated pesticides, polychlorinated biphenyls, polynuclear aromatic hydrocarbons, chlorinated benzenes, and organophosphorous pesticides in water (liquidliquid extraction, GC determination), Atlantic Region Environmental Quality Laboratories, Moncton, NB. ———. 1992a. Biological test method: Acute test for sediment toxicity using marine or estuarine amphipods. Report EPS 1/RM/26. Environment Canada, Environmental Protection, Ottawa, Ontario. ———. 1992b. Biological test method: Fertilization assay using echinoids (sea urchin and sand dollars). Report EPS 1/RM/27, Environment Canada, Conservation and Protection, Ottawa, Ontario. ———. 1992c. Biological test method: Toxicity testing using luminescent bacteria. Report EPS 1/RM/24, Environment Canada, Conservation and Protection, Ottawa, Ontario. ———. 1994. Interim sediment quality assessment values. Environment Canada, Ecosystem Conservation Directorate, Soil and Sediment Quality Section, Guidelines Division. Ottawa, Ontario. ———. 1995a. "Biological Test Method: Test for Growth and Survival in Sediment using Marine or Estuarine Polychaete worms", Environment Canada, Conservation and Protection, Ottawa, Ontario, September 1995. (Draft Manuscript). ———. 1995b. Protocol for the Derivation of Canadian Sediment Quality Guidelines for the Protection of Aquatic Life. Report CCME EPC-98E, Environment Canada, Ottawa, Ontario. ———. 1995c. Users guide to the application form for ocean disposal. Report EPS 1/MA/1, Environment Canada, Marine environment Division, Ottawa, Ontario. ———. 1996. 1996 National Compendium Monitoring at Ocean Disposal Sites. Environment Canada, Ottawa, Ontario. ———. 1998a. Reference Method for Determining Acute Lethality of Sediment to Marine or Estuarine Amphipods. Report EPS1/RM/35, prepared for Method Development and Application Section, Environmental Technology Advancement Center, Environment Canada, Ottawa. ———. 1997. Polynuclear aromatic hydrocarbons and total arochlors in sediments (GC/MSD), Atlantic Region Monitoring and Evaluation Branch, Moncton, NB. 77 ———. 1998b. Pulp and Paper EEM Technical Guidance Document, Environment Canada, Ottawa, Ontario. Fennell, M. and G. van Aggelen. 1997. Biological Assessment of Sediments Collected from Sydney Harbour, Nova Scotia: Site of the 1997 Pollution Gradient Study Conducted by Environment Canada’s National Disposal at Sea Program. (Unpublished Report). Finney, D.J. 1971. Probit Analysis. 3rd edition. Cambridge University Press, Cambridge, Massachusetts. Hurlbert, S. H. 1971. The nonconcept of species diversity: a critique and alternate parameters. Ecology, 52:577–586. Hurlbert, S.H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monog., 54:187–211. Jackman, P. and K.G. Doe. 1997. Toxicity of Sediments from Sydney Harbour, N.S. - Results of the 1997 Pollution Gradient Study. (Unpublished Report), Environment Canada, Atlantic Region, Moncton, NB. Keddy, C., J. C. Greene, and M. A. Bonnell. 1994. A review of whole organism bioassays for assessing the quality of soil, freshwater sediment, and freshwater in Canada. Scientific Series 198. pp125, Environment Canada, Ecosystem Conservation Directorate, Ottawa, Ontario. Long, E.R., L. J. Field, and D. D. MacDonald. 1998. Predicting toxicity in marine sediments with numerical sediment quality guidelines. Env. Tox. Chem., 17(4):714–727. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res., 27:209–220. Margalef, R. 1958. Information theory in ecology. Gen. Syste., 3:36–71. Matheson, R. A. F., G. L. Trider, W. R. Ernst, K. G. Hamilton, and P. A. Hennigar. 1983. Investigation of polynuclear aromatic hydrocarbon contamination of Sydney Harbour, N.S., Environment Canada, Environmental Protection Service, Dartmouth, NS. McIntosh, R.P. 1967. An index of diversity and the relation of certain concepts to diversity. Ecology, 48:392–404. McCullagh, P. and J. A. Nelder. 1989. Generalized linear models. 511pp. Chapman and Hall, Toronto, Ontario. Microbics Corporation. 1992. Microtox. Update Manual. Carlsbad, CA. Moran, T., D. Adams, K. Walker, and B. Zajdlik. 1997. Lambton Industrial Society 1994/1995 St. Clair River Sediment Program Defining Spatial Extent and Environmental Conditions. prepared for the Lambton Industrial Society by Pollutech Enviroquatics Limited, Point Edward, Ontario. Munawar,M., I. F. Munawar, P. Ross, and R. Dermott. 1992. Exploring aquatic ecosystem health: A multi-trophic and an ecosystemic approach. J. Aquatic. Ecosys. Health, 1:237– 252. Norberg-King, T.J.. 1993. A linear interpolation method for sublethal toxicity: The inhibition concentration (ICp) approach (Version 2.0). Tech. Report 03-93 of National Effluent Toxicity Assessment Center, July 1993. USEPA, Environ.Res. Lab.-Duluth, Duluth, Minnesota. PESC (Pacific Environmental Science Centre). 1999a. Metals Analysis in Soils and Sediments. Version 5.0, Issue Date: September, 1999, PESC, North Vancouver, BC. ———. 1999b. Metals Analysis in Water. Version: 3.3, Issue Date: September 1, 1999, PESC, North Vancouver, BC. 78 ———. 1999c. Mercury in Water Analysis. Version 3.7, Issue Date: September, 1999, PESC, North Vancouver, BC. Porebski, L. M. , K. G. Doe, B. A. Zajdlik, D. Lee, P. Pocklington, G. Atkinson, and J. Osborne. 1998. Interpretive guidance for bioassays using pollution gradient studies – Belledune, New Brunswick. EPS Technical Report WM-20, Environment Canada, Marine Environment Division, Ottawa, Ontario. Porebski, L. M., K. G. Doe, B. A. Zajdlik, D. Lee, P. Pocklington, and J. Osborne. 1999. Evaluating techniques for a tiered testing approach to dredged sediment assessment – A study over a metal concentration gradient. Env. Tox. Chem., 18:2600–2610. Schiewe, M.H., E.G. Hawk, D.I. Actor and M.M. Krahn. 1985. Use of bacterial bioluminescence assay to assess toxicity of contaminated marine sediments. Can. J. Fish. Aquat. Sci. 42:1244– 1248. Shannon, C.E. and W. Weaver. 1949. The Mathematical Theory of Communication. The University of Illinois Press, Urbana, IL. Sigg, L. 1987. Surface chemical aspects of the distribution and fate of metal ions in lakes. pp. 319–349, In: Aquatic surface chemistry: Chemical processes at the particle-water interface. Stumm, W. (ed.) , John Wiley and Sons, New York. Simpson, E. H. 1949. Measurement of diversity. Nature, 163:688. Stebbing, A. R. D., V. Dethlefsen, and M. Carr. 1992. Biological effects of contaminants in the North Sea. Results of the ICES/IOC Bremerhaven Workshop. Marine Ecology Progress Series, Volume 91. Stephan, C.E. 1977. "Methods for calculating an LC50" , p. 65-84, In: Aquatic Toxicology and Hazard Evaluation. ASTM STP 634, Mayer, F.L. and J.L. Hamelink (eds.) American Society for Testing and Materials, Philadelphia, PA. Swartz, R.C., 1999. Consensus sediment quality guidelines for polyaromatic hydrocarbon mixtures. SETAC,18:780–787. Underwood, A.J. 1991. Beyond BACI: Experimental designs for detecting human environmental impacts on temporal variations in natural populations. Australian Journal of Marine and Freshwater Research, 42:569–587. USEPA (United States Environmental Protection Agency), 1991. Evaluation of Dredged Material Proposed for Ocean Disposal Testing Manual, EPA 503/8-91/001. February 1991, USEPA, Washington, DC. ———. 1992. An SAB Report: Review of Sediment Criteria Development Methodology for Non-Ionic Organics. Sediment Quality Subcommittee of the Ecological Processes and Effects Committee. EPA-SAB-EPEC-93-002, USEPA, Washington, DC. 79 ———. 1993. Guidance manual bedded sediment bioaccumulation tests. Report EPA/600/R- 93/183. September 1993, USEPA, Washington, DC. ———. 1999. In press. Methods for the derivation of site-specific equilibrium partitioning sediment guidelines (ESGs) for the protection of benthic organisms. USEPA Washington, DC. In Press. USEPA/USACE (United States Environmental Protection Agency/United States Army Corps of Engineers). 1998. Evaluation of dredged material proposed for discharge in waters of the U.S.—testing manual. EPA 823-B-98-004. February 1998. Wilhm, J.L. 1970. Range of diversity index in benthic macroinvertebrate populations. J. Wat. Pollut. Control Fed., 42:221–224. 80 Appendix A Preferred Sample Garman 75™ Coordinates (1996) Sample Site Coordinates 609' 09.5" 46 o 09' 09.7" 60 o 12' 26.8" 610' 12.1" 46 o 10' 10.3" 60 o 12' 13.8" 610' 32.2" 46 o 10' 31.1" 60 o 11' 26.2" 612’ 24.0” 46 o 12' 25.9" 60o 13’ 14.4” 60 o 13' 15.2" 46 o 15' 07.0" 46 o 15' 03.6" 60 o 10' 23.5" 60 o 10' 30.3" Not Previously Used 46 o 15' 21.0" 60 o 33' 35.6" Field Data Table A-1 Sample Location Coordinates Sample Site Number 1 4o 60 o 13' 28.0" 5 4o 60 o 12' 18.8" 6 4o 60 o 12' 26.5" 9 4o 12 St. Ann’s Harbour Reference Site Sample coordinates use North American Datum (1983). The column “preferred sample coordinates” refers to coordinates from the initial site screening survey undertaken in the fall of 1996. Trimble™Sample Site Coordinates 46 o 09' 09.1" 60 o 12' 27.1" 46 o 10' 12.5" 60 o 12' 14.6" 46 o 10' 34.0" 60 o 12' 25.0" 46 o 12' 25.8" 60 o 13' 17.3" 46 o 15' 03.8" 60 o 10' 30.0" 46 o 15' 22.6" 60 o 33' 40.1" Table A-3 Blind Sample Numbers Station Number 1 5 9 6 Ref 1 (old stn 12) St Ann Polychaetes Bioaccumulation Replicate 1 Replicate 2 Replicate 3 Pore water, (PAH/PCBS 66 57 92 36 metals), AVS 46 31 96 9 35 76 72 68 86 Replicate 4 metals, PAH/PCB (SED), TOC, particle size ammonia/sulphide, archives, amphipod, Microtox, echinoid, benthos 80 71 2 100 33 23 34 87 73 6 38 15 8 65 88 77 94 75 40 84 11 48 49 24 Replicate 5 benthos 5 56 70 53 90 89 87 Appendix B Raw Data for Toxicity Tests and Bioaccumulation East Coast Amphipod Biological Toxicity Test Results Table B-1 Sample ID Results of 10-Day Test with Amphiporeia virginiana Percent Survival Sample No. Rep #1 Rep #2 Rep #3 Rep #4 Rep #5 55 90 80 65 70 35 70 15 75 75 60 85 40 75 40 55 0 0 90 75 50 80 85 85 55 75 55 55 0 75 70 60 80 60 95 40 60 0 15 70 70 55 80 85 75 85 85 55 70 5 85 75 60 70 30 90 65 60 0 0 85 65 50 100 70 90 90 85 45 60 0 90 70 55 90 40 85 50 70 0 0 85 70 60 90 95 90 65 60 50 90 0 75 90 60 85 55 90 65 45 5 5 60 80 50 97AT001351 97AT001351 97AT001351 97AT001321 97AT001322 97AT001323 97AT001324 97AT001325 97AT001326 97AT001327 97AT001328 97AT001329 97AT001330 97AT001331 97AT001332 97AT001333 97AT001334 97AT001335 97AT001336 97AT001337 97AT001338 19R1- Martinique 19R2- Martinique 19R3- Martinique 2 6 11 15 23 24 33 34 38 40 48 71 73 75 80 84 87 100 Mean Percent Survival Results of 10-Day Test with Rhepoxynius abronius Percent Survival Sample No. Rep #1 Rep #2 Rep #3 Rep #4 Rep #5 100 100 100 75 100 80 80 95 85 95 60 85 90 100 90 70 90 85 65 95 90 100 100 100 75 100 70 95 70 100 95 60 95 60 85 60 70 70 80 75 85 95 95 100 100 75 95 70 100 95 85 90 70 100 80 100 80 75 80 80 70 85 95 *40 animals were added per jar to all 5 replicates by mistake. 100 100 95 100 92.5 75 100 85 85 95 70 100 70 95 70 80 80 95 70 95 90 97AT001351 100 97AT001351 100 97AT001351 100 97AT001321 90 97AT001322 92.5 97AT001323 95 97AT001324 85 97AT001325 95 97AT001326 95 97AT001327 90 97AT001328 70 97AT001329 95 97AT001330 65 97AT001331 95 97AT001332 95 97AT001333 70 97AT001334 100 97AT001335 90 97AT001336 75 97AT001337 65 97AT001338 100 Table B-2 Sample ID 1R1- Whidby 1R2- Whidby 1R3- Whidby 2 6* 11 15 23 24 33 34 38 40 48 71 73 75 80 84 87 100 Mean Percent Survival SD 16.73 9.35 6.52 14.83 10.61 8.37 13.42 6.52 7.07 8.22 2.24 7.58 12.25 7.58 12.55 9.08 2.24 6.52 12.55 5.70 4.47 81 85 84 72 75 48 69 4 80 76 59 82 45 87 52 58 1 4 78 72 53 SD 2.24 0.00 2.24 11.51 3.79 10.37 9.08 10.95 7.07 2.74 5.48 6.12 12.04 6.12 14.32 4.47 11.40 6.52 4.18 12.25 4.18 99 100 99 83 96 78 92 88 90 93 66 95 73 95 79 73 84 86 71 85 94 Table B-3 Survival Results for the 28-Day Bioaccumulation Test with Macoma nasuta Sample No. Sample ID Total tissue Percent Survival weight (g) 97AT001339 97AT001339 97AT001339 97AT001339 97AT001339 97AT001340 97AT001340 97AT001340 97AT001340 97AT001340 97AT001341 97AT001341 97AT001341 97AT001341 97AT001341 13.12 14.05 12.14 14.80 14.78 9.69 10.46 8.50 12.54 17.20 15.18 15.90 13.96 7.00 9.74 15.01 3.51 12.25 12.71 12.55 13.21 15.64 10.32 11.57 13.23 8.90 7.63 10.09 9.28 9.17 11.19 13.98 12.78 3.14 9.85 3.24 4.36 0 INITIAL-1 INITIAL-2 INITIAL-3 9-1 9-2 9-3 9-4 9-5 31-1 31-2 31-3 31-4 31-5 35-1 35-2 35-3 35-4 35-5 39-1 Control 97AT001345 39-2 Control 97AT001345 39-3 Control 97AT001345 39-4 Control 97AT001345 39-5 Control 97AT001345 97AT001342 97AT001342 97AT001342 97AT001342 97AT001342 97AT001343 97AT001343 97AT001343 97AT001343 97AT001343 97AT001344 97AT001344 97AT001344 97AT001344 97AT001344 46-1 46-2 46-3 46-4 46-5 72-1 72-2 72-3 72-4 72-5 96-1 96-2 96-3 96-4 96-5 Mean Percent Survival/ Rep Rep #1 Rep #2 Rep #3 100.00 88.90 77.80 88.90 55.57 88.90 100.00 100.00 100.00 100.00 44.47 66.67 88.90 33.33 88.90 100.00 88.90 88.90 100.00 66.67 100.00 100.00 83.33 66.67 77.80 88.90 66.67 88.90 100.00 88.90 22.23 66.67 22.23 33.33 0.00 100 100 66.7 66.7 100 66.7 100 100 100 100 66.7 100 100 100 66.7 100 100 100 100 100 100 100 100 0 66.7 66.7 66.7 100 100 100 66.7 100 0 0 0 100 66.7 66.7 100 0 100 100 100 100 100 66.7 0 66.7 0 100 100 66.7 66.7 100 33.3 100 100 100 100 66.7 100 100 66.7 100 66.7 0 66.7 66.7 0 0 100 100 100 100 66.7 100 100 100 100 100 0 100 100 0 100 100 100 100 100 66.7 100 100 50 100 100 100 33.3 100 100 100 0 33.3 0 100 0 SD/ Rep Mean Percent Survival/ Sample 82.23 97.78 64.45 88.89 85.56 86.67 28.89 0.00 19.23 19.23 19.23 50.92 19.23 0.00 0.00 0.00 0.00 38.51 57.74 19.23 57.74 19.23 0.00 19.23 19.23 0.00 33.35 0.00 0.00 28.87 57.74 19.23 19.23 33.35 19.23 0.00 19.23 38.51 33.35 38.51 57.74 0.00 88 89 West Coast Amphipod Biological Toxicity Test Results – Table B-4 Results of 10-day Sediment Assays using Eohaustorius estuarius 12 - 22 Aug 1997 Replicates C B A Treatment #29 - Control - Pachena Bay 100 100 100 100 0 0 0 0 0 0 0.0 % survival %atsurface #29 - Control - Pachena Bay 100 100 100 100 0 0 0 0 0 0 0.0 % survival %atsurface #29 - Control - Pachena Bay 100 100 100 100 0 0 0 0 0 0 0.0 % survival %atsurface Combined Mean for Control Survival 100 100 100 #2 95 100 100 #6 #15 100 0 0 0 0 0 0 0.0 95 0 0 0 0 0 0 0.0 90 90 85 85 95 89 4.2 0 10 0 0 0 2 4.5 95 15 100 10 100 10 100 20 90 90 95 #24 100 100 100 #33 100 90 90 #34 100 100 100 #38 0 100 0 100 0 95 #48 100 90 90 #73 #80 10 45 10 100 10 55 0 90 30 100 0 95 #84 85 100 85 #87 75 90 90 #100 % survival %atsurface % survival %atsurface #11 %survival %atsurface % survival % at surface #23 %survival %atsurface % survival %atsurface % survival %atsurface % survival %atsurface % survival %atsurface #40 %survival % at surface % survival %atsurface #71 %survival %atsurface % survival %atsurface #75 %survival % at surface % survival % at surface % survival %atsurface % survival %atsurface % survival %atsurface #80frozen%survival % at surface 100 10 80 65 60 65 65 67 7.6 0 5 5 0 5 3 2.7 100 0 0 0 0 0 0 0.0 100 0 0 0 0 0 0 0.0 85 0 0 5 0 0 1 2.2 85 0 0 0 0 0 0 0.0 80 90 85 95 80 86 6.5 10 90 0 0 0 0 0 0 0.0 70 90 70 85 90 81 10.2 0 0 0 0 0 0 0.0 65 0 10 5 0 0 3 4.5 75 60 60 75 35 61 16.4 0 70 10 100 0 0 0 0 0 0 0.0 100 0 0 0 0 0 0 0.0 60 0 0 5 0 0 1 2.2 50 75 45 50 65 57 12.5 5 15 20 25 SD Mean E D 0.0 100 100 0.0 100 100 0.0 100 100 0.0 100 2.2 99 95 4.2 96 90 2.2 4.5 99 13 4.2 94 95 2.2 99 95 6.1 90 85 6.7 97 100 5.5 4.5 4 97 10 100 14.8 83 70 12.2 20.8 5.5 4.2 10 68 4 96 0 70 0 95 8.2 94 100 12.4 79 80 7.9 15 10 Results of 10-day Sediment Assays using Eohaustorius washingtonianus Table B-5 29 Jul - 08 Aug 1997 Treatment #14 - Control - Esquimalt Lagoon % survival %atsurface #14 - Control - Esquimalt Lagoon % survival %atsurface #14 - Control - Esquimalt Lagoon % survival %atsurface #6 #33 #48 #80** #100 #2 %survival %atsurface % survival %atsurface #11 %survival %atsurface #15 %survival % at surface #23 %survival %atsurface #24 %survival %atsurface % survival %atsurface #34 %survival %atsurface #38 %survival %atsurface #40 %survival %atsurface % survival %atsurface #71 %survival %atsurface #73 %survival %atsurface #75 %survival %atsurface % survival %atsurface #84 %survival %atsurface #87 %survival %atsurface % survival %atsurface #80frozen%survival %atsurface ** sample seemed to immobilize amphipods; some difficulty telling live from dead 90 Replicates C A SD Mean E D B 6.1 95 95 85 100 95 100 0 0 0 0 0 0 0.0 2.7 97 95 100 95 95 100 0 0 0 0 0 0 0.0 2.2 99 100 100 95 100 100 0 0 0 0 0 0 0.0 4.1 97 Combined Mean for Control Survival 9.1 92 95 85 80 100 4.5 13 15 10 10 20 7.6 93 85 95 85 100 2.2 99 100 100 95 100 27.2 36 25 30 55 0 10.4 53 55 50 45 45 95 95 85 85 95 91 5.5 0 10 0 0 0 2 4.5 100 0 0 0 0 0 0 0.0 65 35 45 70 85 60 20.0 0 5 5 5 25 8 9.7 85 80 65 80 75 77 7.6 10 80 55 50 40 40 53 16.4 10 10 5 0 0 5 5.0 75 85 95 95 75 85 10.0 0 10 10 0 5 5 5.0 100 0 0 0 0 0 0 0.0 55 70 80 65 50 64 11.9 5 10 5 0 0 4 4.2 75 85 85 90 90 85 6.1 0 0 0 0 15 3 6.7 45 65 65 65 55 59 8.9 0 15 5 5 0 5 6.1 100 0 0 0 0 0 0 0.0 65 70 75 85 75 74 7.4 5 0 5 5 15 6 5.5 75 45 75 75 10 56 28.8 0 0 5 10 0 3 4.5 45 50 55 60 60 54 6.5 5 0 0 5 10 4 4.2 70 0 0 5 0 5 2 2.7 95 85 85 95 90 90 5.0 5 0 10 0 5 4 4.2 95 90 75 85 75 84 8.9 5 5 0 0 0 2 2.7 70 0 0 0 0 15 3 6.7 45 25 25 35 35 33 8.4 0 0 5 0 5 2 2.7 East Coast Sea Urchin Biological Toxicity Test Results- Table B-6 Results of Fertilization Inhibition Test with Lytechinus pictus Sample # Sample ID 97AT001321 97AT001322 97AT001323 97AT001324 97AT001325 97AT001326 97AT001327 97AT001328 97AT001329 97AT001330 97AT001331 97AT001332 97AT001333 97AT001334 97AT001335 97AT001336 97AT001337 97AT001338 2 6 11 15 23 24 33 34 38 40 48 71 73 75 80 84 87 100 IC50 (95% Confidence limits) (%) 78.5 (63.7–97.1) > 100 > 100 7.07 (6.01–9.56) 6.04 (5.66–6.72) 35.2 (30.6–41.1) > 100 > 100 8.97 (7.51–10.5) 57.7 (45.7–76.8) 26.6 (-6.0–55.0) 6.17 (5.70–6.67) > 100 > 100 > 100 7.36 (7.07–7.72) 40.6 (33.4–47.7) 28.2 (No CL) IC25 (95% Confidence limits) (%) 52.9 (43.4–65.8) >100 > 100 3.54 (2.99–4.82) 3.03 (2.83–3.35) 15.1 (8.67–20.5) > 100 > 100 4.46 (3.67–5.24) 6.96 (5.72–8.75) 5.09 (4.06–6.44) 3.08 (2.87–3.33) > 100 > 100 > 100 3.68 (3.55–3.83) 9.82 (7.06–17.4) 7.92 (6.12–10.2) West Coast Sea Urchin Biological Toxicity Test Results- Table B-7 Echinoid Fertilization Inhibition Test using the Eccentric Sand Dollar Test Date: Aug 7/97 Site Sample Number _____ 974418-1 974418-2 974418-3 974418-4 974418-5 974418-6 974418-7 974418-8 974418-9 _____ 974418-10 974418-11 974418-12 974418-13 974418-14 974418-15 974418-16 974418-17 974418-18 974418-26 Control 2 6 11 15 23 24 33 34 38 Control 40 48 71 73 75 80 84 87 100 80 (frozen) Percent Fertilization in Replicates Mean at 100% Concentration C B A 91 3 0 59 96 78 94 1 32 93 92 23 9 89 24 31 80 25 1 74 88 97 4 0 75 93 84 97 2 41 94 93 28 15 86 28 40 91 22 2 80 85 94 4 3 70 97 76 95 0 30 97 96 29 12 82 25 28 87 23 0 88 88 Percent Fertilization after Abbott's Correction 3.00 0.58 1.73 8.19 2.08 4.16 1.53 1.00 5.86 2.08 2.08 3.21 3.00 3.51 2.08 6.24 5.57 1.53 1.00 7.02 1.73 100 4 1 72 100 84 101 1 37 100 100 4 1 72 101 84 101 1 37 101 0 94 4 1 68 95 79 95 1 34 95 94 27 12 86 26 33 86 23 1 81 87 91 SD East Coast Polychaete Biological Toxicity Test Results- Table B-8 Sample ID Initial 4 Whitty’s Beach 35 66 68 92 Results of 14-Day Test with Boccardia proboscidea Sample Number 97AT001353 97AT001322 97AT001348 97AT001349 97AT001337 Mean Wt/Treat. Wt/worm (mg) (mg) 0.64 1.401 1.409 1.551 1.583 1.557 0.626 0.558 0.615 0.760 1.233 1.176 1.647 1.546 1.826 0.988 1.361 1.460 1.319 2.091 1.301 1.491 1.694 1.558 1.251 1.828 1.530 1.792 1.441 1.465 SD 0.0857 0.231 0.344 0.370 0.247 0.161 92 Survival (%) N/A N/A N/A N/A 100 100 75 100 100 75 100 75 100 100 100 100 100 100 75 100 100 75 100 75 SD Mean Percent Survival/ treat. N/A N/A N/A N/A 12.50 N/A N/A N/A N/A 93.75 14.43 87.5 0 100 12.5 93.75 14.43 87.5 Results of 14-Day Test with Polydora cornuta Sample Number Wt/worm (mg) 97AT001322 97AT001352 97AT001346 97AT001347 97AT001348 97AT001349 97AT001337 0.109 0.088 0.09 0.326 0.088 0.875 1.541 0.85 0.788 0.705 1.9225 3.1925 2.63 1.029 1.789 1.278 0.868 0.857 0.975 0.788 1.299 0.922 1.038 1.02 0.849 0.41 1.85 0.42 0.348 0.415 1.288 1.08 2.339 1.507 4.989 0.5008 0.898 0.225 0.243 N/A SD Mean Wt/Treat. (mg) 0.104 0.14 0.336 0.952 0.829 2.113 0.194 0.953 0.171 1.026 0.1 0.356 1.609 2.241 0.314 0.466 Table B-9 Sample ID Initial 6 32 Conrad’s Beach 36 57 66 68 92 93 Survival (%) N/A N/A N/A N/A N/A 80 80 100 80 20 100 80 100 100 80 100 80 60 80 100 100 100 80 100 100 80 60 100 100 60 80 100 100 100 100 100 40 40 40 0 SD Mean Percent Survival/ treat. N/A N/A N/A N/A N/A 30.33 N/A N/A N/A N/A N/A 72 10.95 92 16.73 84 8.94 96 20 80 8.94 96 35.78 44 West Coast Polychaete Analyses Table B-10 Boccardia proboscidea in 14-Day Survival and Growth Tests Test period: 7 - 21 August 1997 Test organism age: 20 - 23d old, collected 15 - 18 July 1997 Treatment #4 - Control #6 (Stn. 12) #66 (Stn. 1) #68 (St. Ann’s ) % survival parameter % survival ttl 14d biomass (mg) dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) % survival ttl 14d biomass (mg) dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) % survival ttl 14d biomass (mg) dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) ttl 14d biomass (mg) dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) B A 100 3.77 0.75 0.047 8.0 100 3.45 0.69 0.043 7.3 0.094 60 2.33 0.78 0.049 8.2 80 1.91 0.48 0.027 5.1 0.094 60 1.02 0.34 0.018 3.6 40 2.26 1.13 0.074 12.0 0.094 40 0.99 0.50 0.029 5.3 0 n/a n/a n/a n/a 0.094 Replicates C E D 60 1.91 0.64 0.039 6.8 40 0.96 0.48 0.028 5.1 100 3.01 0.60 0.036 6.4 20 0.05 0.05 -0.003 0.5 40 0.15 0.08 -0.001 0.8 60 1.38 0.46 0.026 4.9 60 1.26 0.42 0.023 4.5 40 2.12 1.06 0.069 11.3 60 1.13 0.38 0.020 4.0 100 2.06 0.41 0.023 4.4 0 n/a n/a n/a n/a 60 1.72 0.57 0.034 6.1 94 SD Mean 28 1.16 0.10 0.007 1.1 80 2.62 0.63 0.038 6.7 23 1.03 0.31 0.022 3.2 52 1.16 0.37 0.020 3.9 11 0.59 0.39 0.028 4.2 52 1.56 0.67 0.041 7.1 42 0.55 0.08 0.006 0.9 40 1.59 0.49 0.029 5.2 Table B-11 Polydora cornuta in 14-Day Survival and Growth Tests Test period: 7 - 21 August 1997 Test organism age: 20-d old, collected 18 July 1997 A parameter Treatment 100 #32 - Control Percent survival Total 14-d biomass (mg) 6.63 1.33 0.090 19.1 0.069 dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) Percent survival #6 (Stn. 12) 100 total 14-d biomass (mg) 4.97 0.99 0.066 14.3 0.069 dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) Percent survival #66 (Stn. 1) 40 total 14-d biomass (mg) 0.11 0.06 -0.001 0.8 0.069 dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) #68 (St. Ann’s) Percent survival 100 total 14-d biomass (mg) 2.87 0.57 0.036 8.3 0.069 dry wt/worm (mg) growth rate (mg/d) growth increase initial wt (mg/worm) Replicates C D B 100 100 100 8.45 1.69 0.116 24.4 8.30 1.66 0.114 23.9 13.09 2.62 0.182 37.7 100 7.74 1.55 0.106 22.3 100 7.80 1.56 0.106 22.5 100 7.01 1.40 0.095 20.2 60 1.42 0.47 0.029 6.8 100 1.21 0.24 0.012 3.5 80 1.52 0.38 0.022 5.5 80 5.53 1.38 0.094 19.9 100 4.47 0.89 0.059 12.9 100 6.87 1.37 0.093 19.8 95 E 100 5.84 1.17 0.078 16.8 100 6.15 1.23 0.083 17.7 100 1.27 0.25 0.013 3.7 80 1.63 0.41 0.024 5.9 SD Mean 0 100 2.81 0.56 0.040 8.1 8.46 1.69 0.116 24.4 0 1.19 0.24 0.017 3.4 100 6.73 1.35 0.091 19.4 26 0.57 0.16 0.011 2.3 76 1.11 0.28 0.015 4.0 11 2.08 0.45 0.032 6.5 92 4.27 0.93 0.061 13.3 Bacterial Photoluminescence Results Table B-12 Bacterial Photoluminscence Results Site 2 6 11 15 23 24 33 34 38 40 48 71 73 75 80 84 87 100 80 (frozen) HS-6 REF 96 Sample number 974418-1 974418-2 974418-3 974418-4 974418-5 974418-6 974418-7 974418-8 974418-9 974418-10 974418-11 974418-12 974418-13 974418-14 974418-15 974418-16 974418-17 974418-18 974418-26 _____ IC50Wet not moisture corrected (%) 0.151 1.07 0.0256 0.690 0.0230 0.410 1.73 0.0221 0.462 0.0353 2.32 0.0349 0.0496 0.0296 0.0255 0.153 0.115 0.0291 0.0224 0.0278 IC50 Dry moisture corrected (%) 0.11 0.83 0.0083 0.23 0.0081 0.14 1.33 0.0081 0.15 0.0140 1.80 0.0147 0.0224 0.0107 0.0103 0.11 0.0828 0.0127 0.0088 _____ Bioaccumulation Results for Macoma nasuta Mean metal concentrations are presented in Table 3 Summary of Macoma nasuta Mean Tissue Metal Levels (g/g dry weight). Table B-13 Summary of Total PAH Tissue Levels (ng/g dry tissue) in Macoma nasuta Site control control control control control 1 1 1 1 1 5 5 5 5 5 6 6 6 6 6 9 9 9 9 12 12 12 12 12 St. Ann’s Harbour St. Ann’s Harbour St. Ann’s Harbour St. Ann’s Harbour St. Ann’s Harbour 97 Subsample 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 1 2 3 4 5 1 2 3 4 5 Response 118.47 127.97 135.72 109.59 114.37 9223.23 9135.14 6538.85 6423.35 9037.66 7050.04 6677.48 6656.63 4812.54 7063.65 4448.33 4641.10 4841.02 3239.62 4606.60 401.91 288.13 396.18 333.96 194.10 147.72 143.26 283.13 147.32 199.40 209.29 226.08 162.48 148.74 Table B-14 Summary of Total PCB Tissue Levels (ng/g dry tissue) in Macoma nasuta Site control control control control control 1 1 1 1 1 5 5 5 5 5 6 6 6 6 6 9 9 9 9 12 12 12 12 12 St. Ann’s Harbour St. Ann’s Harbour St. Ann’s Harbour St. Ann’s Harbour St. Ann’s Harbour 98 Response Subsample 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 1 2 3 4 5 1 2 3 4 5 0.48 0.48 0.48 0.48 0.48 159.5971619 133.5544726 95.96944106 84.19546897 134.1281444 434 385 412 371 443 233 240 272 217 175 0.48 0.48 0.48 0.48 19 17 15 0.48 17 0.48 0.48 0.48 0.48 0.48 Benthic Macroinvertebrate Community Raw Data-Table B-15 Mean Number of Benthic Macroinvertebrate Taxa Site Taxon Aricidea suecica Capitella capitata Capitellidae Cirratulidae Cossura longicirrata Eteone longa Euchone incolor Exogone hebes Harmothoe imbricata Neoleanira tetragona Nephtys ciliata Ninoe nigripes Pherusa plumosa Pholoe minuta Phyllodoce mucosa Polydora quadrilobata Prionospio steenstrupi Sabellidae Scoloplos armiger Spiophanes bombyx Tharyx marioni MarineOligochaete CRUSTACEANS Anonyx sarsi Chiridotea tuftsi Copepoda Diastylis polita Edotea triloba Eudorallopsis deformis Gammarus sp. Leptochirus pinguis Orchomenella pinguis Phoxocephalus holbolli Stenothoe minuta Cylichna gouldi Cerastoderma pinulatum Ilynassa trivittatus Macoma balthica Macoma tenta Margarites groenlandica Mya truncata Nucula delphinodonta Yoldia limatula Edwardsia sp. Echinarachnius parma NEMERTEA NEMATODA Sagitta elegans 99 Station 9 12 6 5 St. Ann's Harbour 1 0 0 0.2 0 0.8 0 0 0 0 0 0 0 0 0 1.4 0 0 0 0.4 0.6 0 0 0 0.4 0 0 0 1 29.4 0.6 0 0 0 0 0.4 0 0.2 0 0.2 0 0 0 0 0.2 0.8 0 0 0 0 0.2 21.4 10.4 13.2 21.6 8.6 0 0.2 0.8 39.2 14 0 0 0.2 2 0.2 0 0 0 0.4 1.4 0 0 0 1.4 0.2 213 2522.4 113 44.8 1.6 0 0 0 0 0.6 0 0 0 1.6 0.6 0 0 0 0 0 0 0 0 0 11.6 0 0 0 0 0 0 2 0 0 0.6 0 0 0.2 0 0 0 0 0 5.2 0 0 0 0 2.2 1.2 0 0 0 0 0.4 0 0.2 0 80.6 7.2 0 0 0 25.4 0.4 0 0 0.2 19.4 0.6 0 0 0 0 0.8 0 0.2 0.4 0 0 0 0 0 24.8 0 0 0 0 0 0.2 0 0 0 106 0.4 0 0 0 0.4 0 0 0 0 2 0 0 0 0 1 12.6 0 0 0 0.2 0.4 0 0 0.2 1.8 4.2 0 0 0 0 0.4 0 0 0 0.8 0 0 0 0 5.4 13.6 0 0 0 2.4 0 0 0.8 0.2 3.8 0.2 0 0 0 0 0.2 0.2 10 1.6 1.4 0.4 0 0 1.2 67.4 67 0 0.2 0 0 0 0 0.2 0 0 0 0 0 0 0 0 21.8 0.2 0 0.2 0 0 0 0 0 0 0 0 0 0.2 0 0 0 0 0 0 0 0 0 0 0.4 0 0 0.2 0 0 0 0 0 0.2 0 1 0 0 Appendix C Rank Correlations Data Table C-1 Summary of Rank Correlations among Sediment Variables with |r|  0.90 Variable 1 Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Variable 2 Acenaphthylene Anthracene Benzo(a)Anthracene Benzo(a)Pyrene Benzo(b)Fluoranthene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Cd Chrysene Cu Dibenzo(a,h)Anthracene Fluoranthene Fluorene Indeno(1,2,3-cd)Pyrene Naphthalene PCB Pb Perylene Phenanthrene Pyrene 1-Methyl-Naphthalene 1-Methyl-Phenanthrene 2,6-Di-Methyl-Naphthalene 2-Methyl-Naphthalene Anthracene Benzo(a)Anthracene Benzo(a)Pyrene Benzo(b)Fluoranthene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Cd Chrysene Cu Dibenzo(a,h)Anthracene Fluoranthene Fluorene Indeno(1,2,3-cd)Pyrene Naphthalene PCB Pb Perylene Phenanthrene Pyrene Rank Correlation 0.9803 0.9873 0.9798 0.9497 0.9434 0.9619 0.9510 0.9550 0.9464 0.9781 0.9307 0.9616 0.9716 0.9882 0.9420 0.9582 0.9513 0.9677 0.9529 0.9893 0.9531 0.9942 0.9781 0.9936 0.9922 0.9936 0.9911 0.9769 0.9725 0.9867 0.9776 0.9798 0.9743 0.9918 0.9487 0.9855 0.9898 0.9897 0.9713 0.9816 0.9745 0.9647 0.9794 0.9948 0.9769 100 Variable 1 Acenaphthylene Acenaphthylene Acenaphthylene Acenaphthylene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene As As As Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene 101 Variable 2 1-Methyl-Naphthalene 1-Methyl-Phenanthrene 2,6-Di-Methyl-Naphthalene 2-Methyl-Naphthalene Benzo(a)Anthracene Benzo(a)Pyrene Benzo(b)Fluoranthene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Cd Chrysene Cu Dibenzo(a,h)Anthracene Fluoranthene Fluorene Indeno(1,2,3-cd)Pyrene Naphthalene PCB Pb Perylene Phenanthrene Pyrene 1-Methyl-Naphthalene 1-Methyl-Phenanthrene 2,6-Di-Methyl-Naphthalene 2-Methyl-Naphthalene Cd Cu Pb Benzo(a)Pyrene Benzo(b)Fluoranthene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Cd Cu Dibenzo(a,h)Anthracene Fluoranthene Fluorene Indeno(1,2,3-cd)Pyrene Naphthalene PCB Pb Perylene Phenanthrene Pyrene 1-Methyl-Naphthalene 1-Methyl-Phenanthrene 2,6-Di-Methyl-Naphthalene 2-Methyl-Naphthalene Rank Correlation 0.9875 0.9860 0.9904 0.9906 0.9981 0.9857 0.9819 0.9927 0.9859 0.9869 0.9563 0.9982 0.9361 0.9910 0.9960 0.9945 0.9802 0.9898 0.9610 0.9575 0.9881 0.9981 0.9856 0.9878 0.9855 0.9913 0.9929 0.9074 0.9254 0.9392 0.9895 0.9864 0.9948 0.9907 0.9917 0.9546 0.9390 0.9953 0.9983 0.9944 0.9862 0.9899 0.9536 0.9577 0.9918 0.9953 0.9878 0.9800 0.9856 0.9849 0.9857 Variable 1 Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Table C-2 Summary of Rank Correlations Among Porewater Variables with |r|  0.90 Variable 1 1-Methyl-Naphthalene 1-Methyl-Naphthalene 1-Methyl-Naphthalene 1-Methyl-Naphthalene 2,3,5-Tri-Methyl-Naphthalene 2,6-Di-Methyl-Naphthalene 2,6-Di-Methyl-Naphthalene 2-Methyl-Naphthalene 2-Methyl-Naphthalene 2-Methyl-Naphthalene 2-Methyl-Naphthalene Acenaphthene Acenaphthylene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Anthracene Benzo(a)Pyrene Variable 2 1-Methyl-Phenanthrene 2-Methyl-Naphthalene Fluorene PCB Acenaphthylene Acenaphthene Salinity Anthracene Fluorene Indeno(1,2,3-cd)Pyrene Phenanthrene Salinity PCB Benzo(a)Anthracene Benzo(a)Pyrene Benzo(b)Fluoranthene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Chrysene Dibenzo(a,h)Anthracene Fluoranthene Fluorene Indeno(1,2,3-cd)Pyrene Phenanthrene Pyrene Eh Benzo(a)Pyrene Benzo(b)Fluoranthene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Chrysene Dibenzo(a,h)Anthracene Fluoranthene Indeno(1,2,3-cd)Pyrene PCB Perylene Phenanthrene Pyrene Benzo(e)Pyrene Rank Correlation -0.9337 0.9860 0.9643 0.9373 0.9919 0.9664 -0.9052 0.9453 0.9948 0.9008 0.9544 -0.9353 0.9330 0.9358 0.9583 0.9642 0.9621 0.9766 0.9413 0.9667 0.9391 0.9241 0.9686 0.9755 0.9943 0.9408 -0.9110 0.9897 0.9877 0.9833 0.9822 0.9944 0.9904 0.9976 0.9988 0.9820 0.9117 0.9700 0.9493 0.9947 0.9985 102 Variable 2 Benzo(e)Pyrene Benzo(k)Fluoranthene Cd Chrysene Cu Dibenzo(a,h)Anthracene Fluoranthene Fluorene Rank Correlation 0.9977 0.9974 0.9251 0.9915 0.9052 0.9956 0.9938 0.9745 Variable 1 Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(a)Pyrene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(b)Fluoranthene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Benzo(k)Fluoranthene Benzo(k)Fluoranthene Benzo(k)Fluoranthene Benzo(k)Fluoranthene Benzo(k)Fluoranthene Benzo(k)Fluoranthene Chrysene Chrysene Chrysene Chrysene Chrysene Chrysene 103 Variable 2 Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Dibenzo(a,h)Anthracene Fluoranthene Fluorene Indeno(1,2,3-cd)Pyrene PCB Perylene Phenanthrene Pyrene Benzo(e)Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Dibenzo(a,h)Anthracene Fluoranthene Fluorene Hg Indeno(1,2,3-cd)Pyrene PCB Perylene Phenanthrene Pyrene Benzo(g,h,i)Perylene Benzo(k)Fluoranthene Chrysene Dibenzo(a,h)Anthracene Fluoranthene Fluorene Hg Indeno(1,2,3-cd)Pyrene Perylene Phenanthrene Pyrene Benzo(k)Fluoranthene Chrysene Dibenzo(a,h)Anthracene Fluoranthene Fluorene Hg Indeno(1,2,3-cd)Pyrene Perylene Phenanthrene Pyrene Chrysene Dibenzo(a,h)Anthracene Fluoranthene Indeno(1,2,3-cd)Pyrene PCB Perylene Phenanthrene Dibenzo(a,h)Anthracene Fluoranthene Fluorene Hg Indeno(1,2,3-cd)Pyrene PCB Rank Correlation 0.9959 0.9978 0.9952 0.9867 0.9195 0.9975 0.9831 0.9569 0.9683 0.9977 0.9989 0.9974 0.9962 0.9933 0.9842 0.9255 0.9940 0.9988 0.9941 0.9509 0.9731 0.9960 0.9978 0.9949 0.9981 0.9906 0.9812 0.9152 0.9816 0.9979 0.9492 0.9685 0.9934 0.9904 0.9981 0.9876 0.9785 0.9285 0.9603 0.9989 0.9384 0.9794 0.9888 0.9962 0.9989 0.9938 0.9912 0.9474 0.9724 0.9525 0.9945 0.9870 0.9242 0.9929 0.9986 0.9893 Variable 1 Chrysene Chrysene Chrysene Cr Cr Dibenzo(a,h)Anthracene Dibenzo(a,h)Anthracene Dibenzo(a,h)Anthracene Dibenzo(a,h)Anthracene Dibenzo(a,h)Anthracene Dibenzo(a,h)Anthracene Fluoranthene Fluoranthene Fluoranthene Fluoranthene Fluorene Fluorene Fluorene Hg Hg Indeno(1,2,3-cd)Pyrene Indeno(1,2,3-cd)Pyrene Indeno(1,2,3-cd)Pyrene PCB PCB Perylene Phenanthrene Zn Ammonia.p 104 Variable 2 Perylene Phenanthrene Pyrene Zn pH Fluoranthene Indeno(1,2,3-cd)Pyrene PCB Perylene Phenanthrene Pyrene Indeno(1,2,3-cd)Pyrene Perylene Phenanthrene Pyrene Indeno(1,2,3-cd)Pyrene Phenanthrene Pyrene Indeno(1,2,3-cd)Pyrene Naphthalene Perylene Phenanthrene Pyrene Pyrene Eh Pyrene Pyrene pH Sulphide Rank Correlation 0.9518 0.9744 0.9958 0.9899 -0.9837 0.9970 0.9878 0.9305 0.9765 0.9498 0.9987 0.9770 0.9765 0.9366 0.9931 0.9381 0.9763 0.9024 0.9702 0.9504 0.9375 0.9817 0.9908 0.9455 -0.9187 0.9712 0.9541 -0.9530 -0.9142