DEVELOPMENT OF A RISK ASSESSMENT TOOL FOR MERCURY IN FISH by Reena Pahal B.Tech., British Columbia’s Institute o f Technology, 2005 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN NATURAL RESOURCES AND ENVIRONMENTAL STUDIES (Environmental Science) UNIVERSITY OF NORTHERN BRITISH COLUMBIA November 2012 © Reena Pahal, 2012 1+1 Library and Archives Canada Bibliotheque et Archives Canada Published Heritage Branch Direction du Patrimoine de I'edition 395 Wellington Street Ottawa ON K1A0N4 Canada 395, rue Wellington Ottawa ON K1A 0N4 Canada Your file Votre reference ISBN: 978-0-494-94091-4 Our file Notre reference ISBN: 978-0-494-94091-4 NOTICE: AVIS: The author has granted a non­ exclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distrbute and sell theses worldwide, for commercial or non­ commercial purposes, in microform, paper, electronic and/or any other formats. L'auteur a accorde une licence non exclusive permettant a la Bibliotheque et Archives Canada de reproduire, publier, archiver, sauvegarder, conserver, transmettre au public par telecommunication ou par I'lnternet, preter, distribuer et vendre des theses partout dans le monde, a des fins commerciales ou autres, sur support microforme, papier, electronique et/ou autres formats. The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. L'auteur conserve la propriete du droit d'auteur et des droits moraux qui protege cette these. Ni la these ni des extraits substantiels de celle-ci ne doivent etre imprimes ou autrement reproduits sans son autorisation. In compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis. Conform em ent a la loi canadienne sur la protection de la vie privee, quelques formulaires secondaires ont ete enleves de cette these. W hile these forms may be included in the document page count, their removal does not represent any loss of content from the thesis. Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant. Canada ABSTRACT Fish can accumulate high levels o f mercury (Hg) and become a human health concern if consumed. The purpose o f this study was to develop a risk assessment tool to determine which water bodies from certain areas in Northern British Columbia contain fish with high Hg concentrations. Raw and published data were collected from Health Canada and Ministry of Environment and amalgamated to form a large data set (3097 fish samples from 34 distinct areas between 1974 and 2000). Fish weight was standardized and a cut­ off point was determined for each species for high Hg levels. This was used to develop a risk assessment tool unique to the study area to identify which species/water body combinations were high in Hg and how fish consumption strategies can be adapted to minimize exposure. Although high Hg levels were widespread, the majority o f contaminated samples were from Pinchi Lake and the Williston Lake area. TABLE OF CONTENTS Abstract ii Table o f Contents iii List o f Tables v List of Figures vi List o f Acronyms vii Acknowledgement viii 1.0 Purpose of the Study 1 2.0 Background and Context o f Study 2.1 Hg in the environment 2.1.1 Hg bioaccumulation in the aquatic food web 2.1.2 Influencing factors for increased Hg concentrations in fish 2.2 Northern British Columbia 2.2.1 Pinchi Lake 2.2.2 Williston Lake 2.3 Benefits of fish consumption 2.4 Dietary concern related to fish consumption 2.5 Benefits versus risks of fish consumption 2.6 Existing risk assessment and management strategies 2.7 Fish consumption advisories 2.8 Risk assessment by FAO/WHO 2.9 Risk management by FAO/WHO 2.10 Risk assessment tool for the Northern Health Authority 1 3 4 7 8 9 10 11 12 15 19 22 30 32 34 3.0 Methodology 3.1 Overall approach 3.1.1 Source of raw data 3.1.2 Use of muscle tissue data only 3.1.3 Use of existing risk assessment tools 3.2 Description o f study area 3.3 Fish collection and Hg analysis 3.4 Statistical analysis o f data 3.4.1 Controlling for length:weight relationships 3.4.2 ANOVA/ANCOVA 3.4.3 Use o f data in the Risk Assessment Tool 35 35 35 35 35 36 36 37 38 39 40 4.0 Results 4.1 Fish species and water bodies included in this study 41 41 iii 4.2 Fish weights and fish lengths 4.3 Mean Hg concentrations in fish by year, location, and species 4.4 Relationship between Hg concentration and fish size 41 49 53 5.0 Development o f the Risk Assessment Tool 5.1 Risk analysis 5.2 Risk assessment 5.2.1 Step 1: Hazard identification 5.2.2 Step 2: Hazard characterization 5.2.3 Step 3: Exposure assessment 5.2.4 Step 4: Risk characterization 5.3 Risk management 5.3.1 Step 1: Identify risk management options 5.3.2 Step 2: Evaluate the options 5.3.3 Step 3: Implementation 5.3.4 Step 4: Monitoring and review 55 55 55 55 55 56 56 56 56 57 57 57 6.0 Risk assessment tool for MeHg levels in fish in Northern BC 6.1 Normal range of Hg in fish in Northern BC 58 58 7.0 Discussion 7.1 High Hg levels in fish from Pinchi Lake and Williston Lake 7.2 Hg levels in fish tissue varied with fish size and sample period 7.3 Hg levels in fish tissue varied with species 7.4 Limitations o f this study 7.5 Development and potential application o f risk assessmenttool 62 63 65 67 68 69 8.0 Conclusion 70 9.0 References 72 10.0 Appendices 79 iv LIST OF TABLES Table 2.1 Forms o f Hg 6 Table 3.1 Number o f fish collected from different water bodies used in this study 37 Table 6.1 Adjusted mean Hg concentration by species and sum o f SD and adjusted means used to determine cut off Hg concentrations for various fish species 59 Table 6.2 Percentage of Hg-contaminated fish by species and water body as estimated using the risk assessment tool 60 Table 6.3 Fish species flagged as potentially contaminated with Hg after adjusting for weight 62 Appendices 79 Table 10.1. Descriptive statistics for fish parameters used in this study Table 10.2 Descriptive statistics for total number o f fish used in this study by water body Table 10.3 Descriptive/tests o f normality for fish parameters used in this study Table 10.4 Tests o f normality for fish weight and Hg concentrations Table 10.5 Square root transformation o f weight of all fish used in this study Table 10.6 Tests of normality for weight o f all fish used in this study Table 10.7 Hg concentration (ppm) with cube root transformation for all fish in this study Table 10.8 Tests o f normality for Hg concentration for all fish used in this study Table 10.9 Interlake differences o f Hg concentration (ppm) in fish tissue, sorted by species Table 10.10 Mean Hg concentration (ppm) in fish tissue by species in various species Table 10.11 Hg concentration in fish tissue collected in various years Table 10.12 Hg concentration in all fish sampled from each location Table 10.13 Mean weight (kg), sorted by fish species Table 10.14 Mean weight (kg) of fish, sorted by location Table 10.15 Hg concentration of fish tissue sorted by year, lake, and fish species Table 10.16 Latin names for fish species used in this study LIST OF FIGURES Figure 2.1 Map o f Williston Lake and the surrounding area 10 Figure 2.2 Generic components of risk analysis 31 Figure 2.3 Generic framework for risk management 33 Figure 4.2.1 Mean weight (kg) of fish caught vs. year caught 42 Figure 4.2.2 Mean weight (kg) of fish caught vs. location 43 Figure 4.2.3 Mean weight (kg) o f fish caught vs. fish type 45 Figure 4.2.4 Mean length (mm) of fish caught vs. year caught 46 Figure 4.2.5 Mean length (mm) o f fish caught vs. location caught 47 Figure 4.2.6 Mean length (mm) of fish caught vs. fish type 48 Figure 4.3.1 Mean Hg concentration (ppm) o f fish caught vs. year caught 49 Figure 4.3.2 Mean Hg concentration (ppm) o f fish caught vs. location caught 51 Figure 4.3.3 Mean Hg concentration (ppm) o f fish caught vs. type of fish 52 Figure 4.4.1 Mean Hg concentration (ppm) in fish tissue vs. fresh weight (gms) o f fish 54 Appendices 79 Figure 10.1 Scatterplot of Hg concentration (ppm) versus weight (kg) for bull trout Figure 10.2 Sample calculation for adjusted Hg concentration (ppm) of fish for the Risk Assessment Tool vi LIST OF ACRONYMS AA - arachidonic acid ALA - alpha-linolenic acid ANCOVA - Analysis of Covariance ANOVA - Analysis of Variance ATSDR - Agency for Toxic Substances and Disease Registry COT - Committee on Toxicity DHA - docosahexaenoic acid DPA - docosapentaenoic acid EPA - eicosapentaenoic acid FAO - Food and Agriculture Association of United Nations FDA - Food and Drug Administration of United States Hg - Mercury HgS - Red sulphide cinnabar MeHg - Methylmercury MWLAP - Ministry of Water, Land, and Air Protection NAS - National Academy of Sciences NMFS -National Marine Fisheries Service PUFA - Polyunsaturated fatty acid SACN - Scientific Advisory Committee on Nutrition SPSS - Statistical Package for the Social Sciences USDA - United States Department of Agriculture U.S. EPA - United States Environmental Protection Agency WHO - World Health Organization ACKNOWLEDGMENT I would like to thank my husband, Jas, for his encouragement, support and knowledge of statistics! I am very grateful to my supervisors, Dr. Laurie Chan and Dr. Michael Rutherford, and my committee member, Dr. Lom a Medd, for watching, directing, and guiding me through each stage o f work. Also, thank you to Dr. Leslie King and Dr. Tina Fraser for being a part o f my defence. Thank you to Ian Baird and everyone at Health Canada and the Ministry o f Environment who helped me to collect all o f my data. 1.0 Purpose of the Study The overall goal of this research was to identify the areas in the Northern Interior of British Columbia where concentrations o f mercury (Hg) in fish are elevated and may pose a risk to fish consumers. Further, this research provided a tool for Public Health professionals in the Northern British Columbia region when assessing the risk associated with fish consumption from various water bodies in the region. Specifically, the following questions were addressed: 1. Which areas in Northern British Columbia have the highest levels o f Hg in fish, and do these levels vary in different species from the same area? 2. Are there similarities in the areas with the highest Hg levels in fish? For instance, are anthropogenic activities associated with the elevated Hg levels? 3. Are there relationships between Hg concentrations in fish and species, age, weight, location, and date o f sampling? 4. Does consumption o f fish from the areas exhibiting high Hg concentrations in fish pose a human health risk? If so, how should Public Health officials respond to this risk? 2.0 Background and Context of the Study Due to its high toxicity to both humans and animals, Hg is one o f the most commonly studied trace elements in the environment (Mousavi et al., 2011). Hg is third (after arsenic and lead) on the 2011 Agency for Toxic Substances and Disease Registry (ATSDR) priority list of 275 hazardous substances, which includes substances that present the most significant potential threats to human health in the United States 1 (ATSDR, 2011). It is released into the environment through both natural and anthropogenic sources and can exist in three forms: elemental, inorganic, and organic (Bhavsar, 2010; Mousavi et al., 2011). In freshwater ecosystems, the organic form (e.g. methylmercury) is predominant and has a high propensity to accumulate in fish tissue through ingestion and absorption (Beyrouty & Chan, 2006). Therefore, dietary consumption of fish and other aquatic animals is a major route o f Hg exposure amongst human and wildlife populations (Mousavi et al., 2011) and the effects of high Hg exposure are well documented (Beyrouty & Chan, 2006). Although methylmercury (MeHg) is produced naturally in the environment (Adriano, 2001), records of it being a potential toxicant and its use in chemical research date as far back as the 1860s (Clarkson, 2002). The commercial production o f organic Hg did not begin until around 1914, when it began to be used as a crop fungicide (Barrett, 2010). Since then, MeHg has come to be known as one o f the m ost hazardous environmental pollutants. Many endemic disasters are attributed to MeHg, such as Minamata disease in Japan and poisoning from the distribution o f wheat seeds dressed with MeHg in Iraq (Mousavi et al., 2011; Legrand et al., 2005). In Canada, Hg pollution surfaced as an issue in 1969 when fish and waterfowl populations within the basins o f Wabigoon and English Rivers in Ontario were found to have elevated Hg levels (Harada et al., 2011). The pollution source was found to be a factory upstream, which used Hg as a catalyst to purify caustic soda. Two indigenous communities along the river (Asubpeeschoseewagong from Grassy Narrows and Wabaseemoong from White Dog) consumed Hg-contaminated fish from the river. Although there is debate about the actual cause o f symptoms, clinical and 2 epidemiological investigations conducted by Harada et al. (2011) found that Minamata Disease-like symptoms were common amongst the population. Follow-up research was conducted in 2002, 2004, and 2010 and the original findings o f these symptoms were reconfirmed (Harada, M., et al., 2011). Even today, Hg concentrations in fish continue to be above safe levels (Kinghom et al., 2007). The amount of Hg released into the atmosphere has increased through other human activities, including coal and municipal waste incineration (Mousavi et al., 2011). Metal mining and smelting, the use o f Hg in gold mining, chlor-alkali production (where Hg is used as an electrode in the electrochemical process of manufacturing chlorine), and bio-medical waste are also anthropogenic sources contributing to increased Hg in the environment (Mousavi et al., 2011). Other examples o f anthropogenic sources o f exposure include: paints and tattoo inks, dental amalgams, barometers, blood pressure monitors, gas regulators, fluorescent bulbs, wall light switches, camera batteries, thermostats, and thermometers (Mousavi et al., 2011). 2.1 Hg in the environment Hg is naturally occurring and is found in air, water and soil; it can be detected almost anywhere in the environment, with normal background levels in sediments usually below 0.1 ppm (ranges between 0.01 to 0.2 ppm). Table 2.1 below summarizes the forms that Hg can exist within the environment. Because Hg exists in many forms, its movement within the environment is influenced by a number o f factors (Adriano, 2001). It cycles naturally through the earth’s crust, atmosphere, oceans, and life forms, with trace amounts in fish, plants, and animals (Rasmussen, 2005). The main ore o f mercury is the red sulphide cinnabar (HgS), which is what is commonly mined. In its gaseous 3 elemental form Hg has an atmospheric lifetime o f six to eighteen months allowing it to be transported around the globe (UNEP, 2008). The main mobilization mechanism for Hg in the environment is through the formation of organic forms. Alkylation is a process which combines inorganic Hg with one or two methyl groups forming monomethylmercury or dimethylmercury (WHO, 2006). Methylation o f Hg is a detoxification process which is performed by microorganisms such as bacterium, fungi and mould; hence the rate of methylation is in part dependent on the abundance o f these organisms (WHO, 2006). 2.1.1 Hg bioaccumulation in the aquatic food web Inorganic Hg, once it has been released by natural and/or anthropogenic sources, enters aquatic environments and accumulates in sediments where it can be transformed into MeHg by sulfate-reducing bacteria under anoxic conditions (Wang et. al, 2012). Anaerobic sulfate-reducing bacteria are the main agents of Hg methylation, and anthropogenic additions o f sulfate are increasing the activities o f these bacteria. These bacteria may methylate Hg in a slow side reaction at 1/1000 the rate of overall sulfate reduction. Sulfate reducers growing at or near redox interfaces may be most important for methylation and Hg contamination o f shallow-water food webs (Fry and Chumchal, 2012 ). Uptake of MeHg from the environment by the lowest organisms o f the food chain plays a key role in MeHg bioaccumulation and biomagnification in biota at higher trophic levels because most o f the Hg that accumulates in species originates from consumption o f organisms at lower trophic levels rather than direct aqueous accumulation. The pathway for MeHg transfer along the food web can be classified as pelagic or benthic according to 4 the foraging habitat. A part of the MeHg in sediments can be taken up by benthic animals directly through gut digestion, or MeHg is also able to enter the water through particulate re-suspension and diffusion, where it can be absorbed by phytoplankton and then biomagnified to potentially harmful concentrations in the food web (Wang et. al, 2012). 5 Table 2.1 Forms ofHg (U.S. EPA, 2007) Types o f Hg Description W here it is found Symptom s from exposure Elem ental or m etallic - N o t carbo n -co n tain in g - A m bient air -R enal toxicity, skin rashes, hypertension, and pulm onary toxicity - S ilver co lo red m etal that ex ists as a thick liquid at room tem perature - T h erm o m eters - N eurological changes (behavioral changes, trem ors, and reduced m uscle coordination) - F lu o rescen t bulbs - D eath, related to respiratory failure - D ental am algam s - P in k disease (sy m p to m s include leg cram ps, irritability, redness /p eelin g o f skin, itching, fever, sw eating, salivating, rashes, sleeplessness) - A v ap o u r in air Organic - P red o m in an tly M eH g - F oods such as fish - E th y lm ercu ry - V accin e preserv ativ es and som e antiseptics - F orm erly used in som e indoor paint - P henylm ercuric A cetate (P M A ) - S edim ent - In utero exp o su re m ay cau se delays in reaching developm ental m ilestones and decrease intelligence - H igh d o ses m ay cau se m ental retardation, reduced m uscle coordination, b lindness, seizures, m uscle w eakness, and inability to speak - E ffects in ad u lt hum ans include kidney dam age and digestive tract problem s -C hronic exposure is linked to elevated blood pressure, increased risk o f heart attack, heart p alpitations, hand trem ors, im paired hearing, dizziness, and staggering -P ink disease Inorganic - N o t carb o n -co n tain in g - N o n -elem en tal form s o f inorganic H g, including m ercuric chloride, m ercuric acetate, m ercuric sulfide, etc. - C o m m ercially available products - C an be toxic to kidneys, stom ach, intestines - C an lead to increased blood pressure - M edicinal hom eopathic herbal rem edies - P o ssib le em bryotoxic effects including increased rates o f m iscarriage and stillbirth - L o w exposure from in d o o r air 6 2.1.2 Influencingfactors fo r increased Hg concentrations in fish The relationship between elevated Hg concentrations in fish in newly flooded areas is well documented. The flooding of vegetation and terrestrial soils through natural and anthropogenic processes contributes to elevated Hg concentrations in the food web o f flooded environments (Mast and Krabbenhoft, 2010). It has been proven that reservoir formation often leads to elevated Hg levels in fish relative to pre-impoundment concentrations, even in cases where no point source discharges o f Hg are evident (Mast and Krabbenhoft, 2010). In newly created reservoirs, the initial flooding o f organic-rich soils can result in elevated Hg concentrations in fish for up to 10 to 20 years later (Bodaly et al., 2007). Hg accumulation may also be elevated in established reservoirs that experience annual water-level fluctuations related to water storage, power generation, or flood control (Mast and Krabbenhoft, 2010). The source o f Hg in reservoirs is likely a redistribution of the element from materials already in the lake or river prior to flooding (Mast and Krabbenhoft, 2010). In situations where the reservoir is reflooded, declining water levels may allow the growth o f vegetation on exposed littoral areas, which then become a new carbon source when the sediments are reflooded, causing an increase in microbial activity and MeHg production. An alternate explanation is that drying o f soils and sediments results in oxidation of reduced sulfur to sulfate which stimulates sulfatereducing bacteria and MeHg production when rewetted (Mast and Krabbenhoft, 2010). The magnitude o f increases in Hg levels in the environment, and in turn, in fish depends on many factors including the area o f land and vegetation inundated, water temperature, pH, alkalinity, sulfate, dissolved organic carbon and the age and retention time o f the reservoir (Mast and Krabbenhoft, 2010). 7 The correlation between the rising concentration o f Hg in fish tissue and the size and age of the fish is well documented. The levels of increase depend on trophic status and diet; the lowest levels are in aquatic plants, intermediate in invertebrates and highest in fish, and piscivorous mammals and birds (Storelli et al., 2007). “Larger, older, and higher-trophic-level fish species generally have higher MeHg tissue residues than smaller and younger organisms from lower trophic levels. Concentrations in top predator fish can be up to 10 million times higher than those in water” (Mahaffey et al., 2011). Studies also show differences in Hg concentrations in pelagic and benthic species; animals living in close association with sediments (in which they bury and from where they feed) are eventually more exposed to sediment-associated contamination than other fish (Storelli et al., 2007). Almost all o f the Hg found in biological systems has been absorbed in the form of MeHg and all freshwater fish in North America, and perhaps in the world, have at least trace levels of Hg in their tissues (Bhavsar, 2010). MeHg can bioaccumulate and biomagnify within aquatic food webs and is highly absorbable to both fish and human consumers via ingestion (95 to 100%) compared to inorganic Hg (5 to 10%) (Storelli et al., 2007; Chan et al., 2003). 2.2 Northern British Columbia Northern British Columbia’s geographic area is approximately 500,000 square kilometers, which comprises more than half o f the province (TourismBC, 2008). This area has many rivers and lakes and is known for its freshwater and saltwater fishing. Hg levels in fish have been a source o f concern in some regions o f Northern British Columbia both in the past and at the present time. Higher than normal levels have been attributed to anthropocentric activities such as the building o f reservoirs and mining 8 close to lakes. The two water bodies that have received the most attention for high Hg levels in Northern British Columbia are Pinchi Lake and Williston Reservoir. 2.2.1 Pinchi Lake The area along the Pinchi fault in central British Columbia is a prime example o f a natural Hg source found and exploited by humans leading to elevated Hg levels in the lake (Weech et al., 2004). A portion o f the northern shore was mined for Hg from 1940 to 1944, at which time the mine was closed and all structures were subsequently demolished. The mine was redeveloped and operated once again from 1968 to 1975. Waste ore was routinely deposited directly into Pinchi Lake; since then, relatively high concentrations of Hg have been observed in the water, sediments, and fish in the area (Weech et al., 2004). There has been media attention on the elevated Hg levels o f Pinchi Lake’s fish, as the Tl’azt’en Nation, a group o f Carrier Indians who live north o f Fort St James, claim that it has affected the health o f the majority o f their population o f 1200 (The Province, 2003). In February o f 2010, a consulting agency prepared a document for Teck Metals (also known as Cominco) entitled “Human Health Risk Assessment of the Pinchi Mine and Pinchi Lake Area.” This document included a closure plan for the mine and concluded that “post-closure environmental conditions and land uses at the mine site described in the Closure Plan should result in acceptable risks to human health for on-site receptors” (Wilson, 2010). An assessment conducted at a later date to ensure that the post-closure conditions are indeed acceptable would be ideal. 9 2.2.2 Williston Lake The Williston Lake Reservoir (Figure 2.1) is located in close proximity to Hudson’s Hope, B.C.; it was created in 1968 by the impoundment of the Peace River in the Peace Canyon for the purpose o f hydroelectric generation (Stockner, 2005). This reservoir is a product o f the W.A.C. Bennett Dam, which flooded land surrounding the Peace, Parsnip, and Finlay rivers (creating the present three reaches) during the late 1960s through to the early 1970s. This is British Columbia’s largest reservoir, with a surface area o f nearly 178,000 hectares and a catchment area close to 70,000 square kilometers (Baker et al., 2000). The lake is used by the sports fishing industry as well as a First Nations community, Tsay Keh Dene, located at the mouth o f the Finlay Reach. Figure 2.1 Map of Williston Lake and the surrounding area (ILEC, 2008) Plotey Wiver, A Fort Ware | . WHfistoa laktf Riv«r\ f mate? Forth? (SU* §5) f P « c« Canyon Oom Fort St, John gf aytet t X^*Hud?on'? Hop* V.AXX Bennett Dim Scale km too 10 2.3 Benefits o f fish consumption The importance o f fish consumption for good health and nutrition is well accepted; it has provided humanity with an important food source for thousands o f years. Fish are a source o f many vitamins, including niacin, vitamins B12, D, and A. Further, fish provide a dietary source o f other nutrients including selenium, iodine, fluoride, calcium, copper, choline, taurine and zinc (SACN/COT, 2004; Karagas et al., 2012; FAO/WHO, 2011). The many health benefits o f fish are partly due to the high concentrations of n-3 polyunsaturated fatty acids (n-3 PUFAs) present in many species (Mahaffey et al., 2011). The n-3 PUFAs that are particularly important in human nutrition include alphalinolenic acid (ALA), eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and docosahexaenoic acid (DHA). The two fatty acids that are especially important for human neurological development are EPA and DHA (Mahaffey et al., 2011). These essential fatty acids play important roles in cell membrane formation, integrity, and functions; the functioning of the brain, retina, liver, kidney, adrenal glands, and gonads; and local hormone production for the regulation of blood pressure and immune and inflammatory responses (Mahaffey et al., 2011). Maternal intake of fish has been observed to be valuable to fulfill fetal requirements. DHA and arachidonic acid (AA), an omega-6 PUFA, are essential for the development of the central nervous system in mammals (SACN/COT, 2004). During the last trimester of pregnancy, fetal requirements for DHA and AA are very high due to the rapid synthesis of brain tissue. The main sources o f the DHA and AA that accumulate in the brain are drawn from maternal circulation during pregnancy and through breast milk 11 for newborns (Mahaffey et al., 2011; SACN/COT, 2004). In pre-term and low-birthweight babies, DHA deficiency has been related to visual impairment and delayed cognitive development. Also, there is some evidence that increased maternal intake o f fish or fish oil supplements may prolong gestation, minimizing preterm delivery and low birth weight (SACN/COT, 2004). Multiple observational studies have monitored DHA levels in maternal blood during pregnancy, in umbilical cord blood during delivery or o f maternal fish consumption during pregnancy. These studies demonstrate independent beneficial associations o f DHA levels with more optimal neurodevelopmental outcomes in offspring, such as better behavioural attention scores, visual recognition memory and language comprehension in infancy and childhood (FAO/WHO, 2011). The ingestion o f fish or fish oils has been associated with an array o f health benefits including improvement of blood lipid profiles, decreased risk of heart disease, lowered blood pressure, improvement in rheumatoid arthritis, enhanced eye and brain development in early life, prevention o f macular degeneration, less risk o f colitis and type 2 diabetes, and improvement in neurological and psychological disorders such as depression, schizophrenia and Parkinson’s Disease (Ginsberg and Toal, 2008). A number of studies have shown strong evidence that fish or fish oil consumption reduces all-cause mortality and various cardiovascular disease outcomes (FAO/WHO, 2011). 2.4 Dietary concern related to fish consumption Monomethylmercury is a neurotoxic species which bioaccumulates in fish tissue and is a principal Hg-related human health concern today (Mousavi et al., 2011). As it is readily formed in water and remains in the water column, fish take in MeHg by ingestion 12 of contaminated prey and particles or assimilate the compound through their gills during respiration. The toxicity o f low concentrations o f Hg on most aquatic organisms can result in decreased growth rate, reproduction and overall ability to survive (Mousavi et al., 2011). Hg and MeHg compounds usually affect the nervous system, the kidneys, and the developing fetus. Fetal brain Hg levels tend to be 5 to 7 times higher than in the mother’s blood, with the developing central nervous system being of highest concern (WHO, 2006; Karagas, 2012). Hg toxicity can also affect fetal respiratory, cardiovascular, gastrointestinal, hematological, immune and reproductive systems. There are several factors that determine the presence and severity o f adverse health effects such as the chemical form o f the Hg, the dose, the age or developmental stage o f the exposed individual, the duration o f exposure, and the route o f exposure (WHO, 2006). Three routes of exposure are dermal contact, inhalation, and ingestion, with ingestion being the most common and hazardous. Almost all Hg in fish is in the form o f MeHg, which has a high affinity for proteins in fish muscle (Melwani et al., 2009). All humans are exposed to some level o f Hg, and fish is the primary source o f Hg exposure among the general population (Bhavsar et al., 2010). Because o f this, many governments provide dietary advice to limit fish consumption where there are elevated Hg levels (WHO, 2006). The W orld Health Organization and the Food and Agriculture Organization o f the United Nations recommend a maximum o f 0.5 ppm in non-predatory fish and 1 ppm in predatory fish. The United States Food and Drug Administration set a maximum level o f 1 ppm in fish, 13 shellfish and aquatic animals. The European Community allows 0.5 ppm in fishery products and Japan allows up to 0.3 ppm. Health Canada has set a guideline o f 0.5 ppm for the Hg level in commercial fish; however, certain predatory fish species sold in Canada may contain Hg levels that are higher than this. In 2007, Health Canada issued a new standard which limits acceptable Hg content in predatory fish to 1.0 ppm (Health Canada, 2007b). Such predatory species include fresh/frozen tuna (Thunnus spp.), shark (Selachimorpha spp.), swordfish (Xiaphis gladius), marlin (Makaira spp.), orange roughy (Hoplostethus atlanticus), and escolar (Lepidocybium flavobrunneum). There is no limit on the consumption o f fish such as salmon (Salmo salar (Atlantic); Oncorhynchus spp. (Pacific)), cod (Gadus spp.), pollock (Pollachiuspollachius), sole (Solea spp.), shrimp (Pandalus borealis), mussels (Mytilus edulis), scallops (Pecten maximus), and canned light tuna (Thunnus spp.). There is also no limit on the consumption o f the species used in this study (including bull trout, dolly varden, and lake trout, for which there are current advisories in Northern BC), nor do the guideline limits differ according to where the fish is caught (for example, inland/offshore, which province/area, etc.). Many communities, especially First Nations, rely on fish intake as a daily component of their meals and nutrient intake and are likely at risk for chronic exposure to MeHg (Chan et al., 2011). It is important to balance the risks and benefits o f fish consumption when nutritional, social, cultural, and economic benefits are concerned (Chan et al., 2011). In many communities, researchers are aware that MeHg is present in the area; however, there is a lack o f comprehensive data collection and analysis. Collecting and analyzing data can be a lengthy and expensive process; limited human and 14 financial resources make it difficult to study each community in depth. That being said, there is a current and ongoing “First Nations Food, Nutrition, and Environment Study” which analyzes food and water consumption for one hundred randomly selected First Nations communities across Canada (in British Columbia, Manitoba, and Ontario) (FNFES, 2012). This project is being conducted by the University o f Northern British Columbia and has principal investigators from the University o f Ottawa, University o f Montreal, and the Assembly o f First Nations, as well as co-investigators from Health Canada. This study also reports average total Hg in hair concentrations and conducts household interviews about traditional foods consumed in the past year (FNFES, 2012). 2.5 Benefits versus risks o f fish 'consumption Whether the benefits o f fish consumption outweigh the risks of MeHg exposure has been a controversial issue for quite some time. The most important factors in the debate are the MeHg exposure levels in fish being consumed, and the level o f risk associated with MeHg. When taking different factors into consideration, the recommendations can be quite confusing. For instance, the Food and Drug Administration o f the United States (FDA) issued consumption advice in 2004, but only for children and pregnant and/or nursing women. Health Canada (2011) recommends 150g/week as a general guideline, 150g/month for pregnant women, 125g/month for children 5-11 years o f age, and 75g/month for children 1-4 years o f age (for nonpredatory fish). To add to the confusion, the American Heart Association (2010) recommends at least 2 fish meals per week (Turyk, 2012). Two studies from the Seychelles and Faroe islands show contrary assessments despite having similar populations with high per capita consumption of fish and MeHg 15 body burdens higher than in the United States (Ginsberg and Toal, 2008). The Faroe study showed significant neurodevelopmental deficits at birth and into the early school years whereas the Seychelles study, at similar MeHg exposure levels, showed no evidence of harm. The National Academy o f Sciences (NAS) held a peer review panel to determine the reasons for the conflicting results o f the two studies and found that there were four main differences. The Faroese study used the umbilical cord to test for Hg concentration while the Seychellois survey used maternal hair; the Faroese study used domain-specific tests and the Seychellois study used globally-accepted tests (which may not have been sensitive enough to the region); the Faroese children were evaluated at age 7 and the Seychellois at age 5.5 (a less sensitive age for neuropsychological tests); the Faroese eat whale meat, which can significantly raise the concentration o f Hg in the body in a short amount o f time. After taking the above differences into account, NAS determined that the study from the Faroe Islands, even though they eat whale meat, fulfilled the most criteria for use o f an epidemiology study in risk assessment (Jacobson, 2001). This was mostly due to the fact that the Seychelles study may have lacked sufficient power to detect the relatively small effect sizes computed for the Faroe Islands data (Price et al., 2007). Based on these two studies and an additional one from New Zealand, a National Academy of Sciences report in 2000 (NRC, 2000) concluded that MeHg in fish is an important public health risk and a dose-response analysis for neurodevelopmental effects was developed. The U.S. EPA used this to derive a reference dose (RfD) o f 0.1 pg/kg body weight/day (U.S. EPA, 2011). 16 Many fish consumption advisories intend to educate the public on balancing the risks and benefits o f fish consumption; however, messages in the media that emphasize fish benefits have created confusion about the need for caution (Hobson, 2006). In some cases, warnings about MeHg levels in fish portray overly negative messages that cause individuals to completely avoid fish (Cohen et al., 2005; Oken et al., 2003). In 2001, when the U.S. FDA recommended “that women o f child bearing age should avoid consuming specific long-lived predatory fish high in Hg and limit fish and shellfish meals, pregnant women in eastern Massachusetts decreased their total fish consumption, resulting in an estimated decline o f 17%” (Turyk, 2012). On the other hand, some advocacy groups have recommended that pregnant women exceed federal fish consumption guidelines (Couzin, 2007). There are many factors that need to be considered when estimating safe levels o f MeHg intake from eating fish, making it a very difficult task. Factors such as differing sensitivity to MeHg amongst various populations; varying types, amounts, and frequency o f seafood consumed; and the differences in Hg concentration o f species all need to be acknowledged. Further, there is disagreement about the reliability, variability, and interpretations of existing data (Mahaffey et al., 2011). However, there are a few points that all agree with: pregnant women and young children are the most sensitive groups, MeHg is neurotoxic, and seafood is the primary dietary source. Moreover, the many health benefits associated with seafood consumption, especially during pregnancy and early life, are well known (Mahaffey et al., 2011). Mahaffey et al. (2011) provides a comprehensive review o f many risk-benefit considerations of fish consumption on child development that have been published in 17 recent years. This work sheds light on significant advances that have been made in understanding the toxicology and epidemiology o f MeHg exposures, as well as the nutritional benefits of n-3 PUFAs. However, a number of knowledge gaps still remain, including the need for much more information on the quantities o f n-3 PUFAs that can be synthesized from ALA through maternal metabolism. Many studies have considered the association between fish intake and child development at relatively low exposure levels, but unfortunately, not all studies provide detailed seafood consumption results (Mahaffey et al., 2011). This makes conducting a quantitative benefit/risk assessment o f DHA intake and MeHg exposure challenging because seafood consumption remains a major determining factor. It is made even more complex when additional factors are taken into account, such as exposure to lipophilic organic contaminants (such as PCBs, as they also tend to accumulate in predatory fish), other nutrients, or variability in individual body weights (Mahaffey et al., 2011). The World Health Organization (WHO) and the Food and Agriculture Organization of the United Nations (FAO) have come together to formulate fish consumption advice. Their most recent document, “Report o f the Joint FAO/WHO Expert Consultation on the Risks and Benefits o f Fish Consumption” was released in 2011. This report concludes that the consumption of fish provides energy, protein, and a range o f other nutrients, as well as reducing the risk o f mortality from coronary heart disease. There is an absence o f probable or convincing evidence o f risk o f coronary heart disease associated with MeHg and when comparing the benefits o f n-3 PUFAs with the risks of MeHg among women o f childbearing age, maternal fish consumption lowers the risk o f suboptimal neurodevelopment in their offspring (FAO/WHO, 2011). However, the 18 available data are currently insufficient to derive a quantitative framework for assessing the health risks and health benefits o f eating fish for infants, young children and adolescents. This document also includes a series o f steps that are recom m ended for member states/countries to better assess and manage the risks and benefits o f fish consumption and more effectively communicate with their citizens. This includes acknowledging that fish is an important food source and is part o f the cultural traditions of many peoples, emphasizing the benefits o f fish consumption on reducing mortality from coronary heart disease (and the risks o f mortality from coronary heart disease associated with not eating fish) for the general adult population, and emphasizing the net neurodevelopmental benefits to offspring o f women o f childbearing age who consume fish, particularly pregnant women and nursing mothers. WHO/FAO (2011) also recommended that jurisdictions develop, maintain and improve existing databases on specific nutrients and contaminants, particularly MeHg and dioxins, in fish consumed in their region; and to develop and evaluate risk management and communication strategies that both minimize risks and maximize benefits from fish consumption. WHO/FAO (2011) considered seven other existing international risk-benefit activities when creating a matrix comparing levels o f the n-3 PUFAs, DHA, and EPA with levels o f total Hg and dioxins developed using existing data. The matrix categorized fish species by one of four levels o f each o f these substances. 2.6 Existing risk assessment and management strategies It is important to understand and complete a risk assessment in order to know the true risk of consuming certain foods. Risk assessment refers to a process that characterizes the degree and nature o f a given risk (Health Canada, 2007a). Through risk 19 assessment, it can be determined if there is a need for risk management, which is prevention and control employed to reduce risk (Health Canada, 2007a). Risk management strategies have been employed in the past to reduce the risk o f unacceptable exposures to Hg and the need for one was first identified by Health Canada in the late 1960s, when a standard was developed for Hg in fish (Health Canada, 2007a). “In their everyday lives, people face hazards and must make individual decisions about the risks they face. However, the public depends upon the government to provide both safeguards and warnings about potential hazards, and to regulate or remediate where needed to reduce exposure to these hazards” (Burger, 2005). A risk assessment guide is needed to provide guidance to risk managers so that they can better understand the risk posed by MeHg in fish. This will provide them with the knowledge needed to assist in developing appropriate cost-effective intervention strategies in which the risk o f fish consumption can be minimized while the benefits can be maximized. The risk-benefit assessment used for the purposes of this thesis is the Codex Alimentarius risk assessment paradigm by WHO/FAO and will be discussed in detail below. However, there are many other international risk-benefit strategies that have been developed. The Benefit-Risk Analysis o f Foods (BRAFO) has gained much attention in Europe, as well as worldwide (ILSI, 2012). This system uses a tiered approach to assess the risks and benefits o f changing from the reference scenario to an alternative, resulting in a statement about which scenario is preferred in terms of health effects. In Tier 1, each risk and benefit is assessed independently, often using standard screening methods (but more refined methods may provide the benefit o f needing to proceed to Tier 2). Tier 1 comprises a separate risk assessment and a separate benefit assessment. In Tier 2, risks 20 and benefits are compared in a qualitative way without the use o f a common metric. In Tier 3, risks and benefits are integrated quantitatively in a common metric, by a deterministic approach. In Tier 4, risks and benefits are integrated quantitatively in a common metric by a probabilistic approach. Another tiered approach was developed by the Scientific Committee o f the European Food Safety Authority (EFSA) which focuses on human health risks and human health benefits, and does not address social, economic and other considerations (EFSA Journal, 2010). This approach is very similar to both the Codex and the BRAFO approach and incorporates three steps: initial assessment (risks versus benefits), refined assessment (quantitative), and the comparison o f risks and benefits (using a composite metric such as DALYs or QALYs). This approach identifies that separate consideration is needed where differences in the sensitivity to the agent under consideration exist or are assumed to exist in specific subpopulations. DALY (Disability Adjusted Life Years) and QALY (Quality Adjusted Life Years) are both ways to assess the burden o f disease attributable to an environmental factor. They are technically similar in that they both express health in time (life years) and give a weight to years lived with a disease, capturing both quality and quantity o f life in one indicator. However, the DALY approach also gives an indication o f the potential number of healthy life years lost due to premature mortality or morbidity and are estimated for particular diseases, instead o f a health state. Although QALYs and DALYs stem from the same broad conceptual framework, they are not interchangeable, as they are partly based on different assumptions and different methodologies (Sassi, 2006). 21 Using these many risk-benefit assessments, government agencies around the world develop fish consumption advice for the public. This often results in fish consumption advisories, the development o f which is also a process which builds upon existing literature and advisories issued by other governments. 2.7 Fish consumption advisories As o f 2008, all 50 states and the District o f Columbia in the United States o f America have issued fish consumption advisories to alert residents of consumption restrictions on certain species from local lakes and rivers (Lando and Zhang, 2011). State advisories vary in their specifics due to differences in fish species and types o f pollutants in local waterbodies. In 2008, 80% o f these advisories, including at least one in every state, were issued in part due to Hg contamination (Lando and Zhang, 2011). The FDA issued national fish consumption advisories in 2001 and 2004, targeting women of childbearing age and households with young children. The efficacy o f these advisories has been evaluated in different ways. Lando and Zhang (2011) examined changes in consumer awareness of Hg contamination in fish and their knowledge o f the information contained in the national advisories by using nationally representative surveys in 2001 and 2006. They tried to test whether the targeted groups in the national advisories were more aware o f the information contained in the advisories in 2006 than they were in 2001. The results indicated that the United States’ population’s awareness of Hg as a potential toxicant in fish increased from 69% to 80% between 2001 and 2006. The percent o f those who could name a targeted at-risk group or fish listed in the national advisories increased and there was also an increase in the mean index score in measuring awareness o f Hg in fish and knowledge o f the information contained in the advisories. 22 Media attention and many federal, state, and local education activities surrounding the national and state fish advisories aided in alerting the public about the potential problem o f Hg in fish. Despite this increase, overall knowledge about the information contained in the national advisories remains low (Lando and Zhang, 2011). Although women in general had greater gains in their level o f awareness and knowledge between 2001 and 2006 than their male counterparts, women o f childbearing age did not have greater awareness or knowledge than the rest o f the population groups. However, adults that had children five years of age or younger in their households had greater awareness and knowledge than those who did not. It was also found that Caucasians, older adults, and highly educated consumers had higher awareness and knowledge index scores than ethnic minorities, younger adults, and less educated consumers. Further, those who lived within easy access to fresh fish and fishing had higher awareness and knowledge index scores (Lando and Zhang, 2011). Another study conducted to assess the efficacy o f the FDA’s 2001 advisory found that it significantly reduced fish consumption amongst the population (Shimshack et al., 2010). This study had a rich data set with household-level consumer panel data from Information Resources, Inc. (IRI). The data set included every packaged supermarket fish purchase from a panel o f nearly 15,000 households in the year before the advisory and the 2 years after the advisory (2000—2002). The consumption data were combined with detailed information on more than 5300 unique products comprising over 50 species. Home fish consumption was then translated into household Hg and omega-3 intakes based on measurements reported in the scientific literature and extensive USD A (United States Department of Agriculture) testing. The empirical findings o f this study showed 23 that Hg intakes fell 17.1% in response to the advisory on average, with the reduction being concentrated among college-educated, high fish-consuming households. However, at-risk consumers’ omega-3 intakes from this food source also fell 21.4%. It appears that the recommendation to continue consuming healthful levels o f seafood and to substitute towards lower Hg fish was not heeded on average (Shimshack et al., 2010). Burger and Gochfeld (2008) interviewed 174 individuals (including students, maintenance staff, and faculty) at Rutgers University in New Jersey to assess the degree of knowledge about the benefits and risks o f fish in relation to ethnicity and the degree o f knowledge. Their study found that people are generally more aware of the benefits o f fish consumption than the risks, and they have more specific information about those benefits than they do about the risks (Burger and Gochfeld, 2008). Further, there were ethnic differences in knowledge about: advisories or benefits, specific information about the risks and benefits, and that some fish are better or worse with respect to the risks from chemicals. Caucasians related more specific information about the risks and benefits than minorities (Burger and Gochfeld, 2008). A review o f literature reveals that awareness o f warnings is sometimes ethnically related; even though minorities tend to consume more fish than their Caucasian counterparts, their knowledge o f advisories is often considerably less (Burger and Gochfeld, 2006). Reasons for the lack o f knowledge may be due to lower income and education; whereas the higher consumption may be culturally related. A pilot study was conducted in Philadelphia on fish consumption and advisory awareness among the Asian community by distributing questionnaires (Perez et al., 2012). This study found that the concept that fish consumption can have both harmful and beneficial effects is a difficult 24 one for populations that rely on seafood as a major dietary and cultural component. Study data were used to evaluate the efficacy o f state-issued advisories and it was found that while advisory awareness levels among study participants were greater than previously observed in Asian-American populations, consumption levels remained high. However, the sample size was quite small (n=34) and represented a very limited sample o f the Philadelphia Asian-American population. DeWeese et al. (2009) reported on the efficacy o f lake-specific, risk-based, culturally sensitive fish consumption advice for tribes in the Great Lakes Region. Areaspecific advisory maps, which were a combination of text and graphics and provided consumption advice as well as information on health benefits o f consuming fish (in particular Ogaa/Walleye), were distributed to tribes in Wisconsin, Minnesota, and Michigan. A behavioural intervention program was developed and the efficacy o f it was assessed using surveys of tribal fish harvesters and women o f childbearing age. Fifty-one families from 10 tribes recorded their fish consumption during the study. The intervention involved dissemination o f the advisory maps to tribal leaders, fish harvesters, women o f childbearing age, children, and elders, as well as the broader tribal population. There were oral presentations which included detailed training on use of the maps, general information about the adverse health effects o f Hg exposure, and information about how map-based consumption advice was developed. After the intervention, concern about Hg increased significantly among all harvesters, but not among women of childbearing age. Although nearly 100% o f Wisconsin tribal harvesters and over 90% o f tribal harvesters in Minnesota and Michigan surveyed found the advisory maps to be very or somewhat helpful, there was no significant increase in the 25 number of harvesters who used maps to make choices about which lakes to harvest. The intervention effort toward women o f childbearing age resulted in an increase o f awareness and concern but not in behavioral changes. Overall, Ogaa harvest in Wisconsin, Michigan, and Minnesota increased during and after the map-based intervention program. Consumers may limit fish consumption or choose among different kinds o f fish based on consumption advisories and media warnings; however, there is a rich literature indicating that this is not always the case (Burger and Gochfeld, 2006). Burger and Gochfeld (2006) reviewed the issuance o f fish consumption advisories, compared angler compliance and knowledge about such advisories, and proposed a framework for information needs necessary to integrate several aspects of fishing, fish consumption, and risk. They found that public health officials need to take a multi-faceted approach to managing the risk that includes cultural sensitivity and audience-specific positive information. They also suggest that more graphics and tables be added to the advisories to make the information easier to understand and absorb. “Whether and how a person responds to consumption advisories depends upon their level o f trust in the conveyor o f risk information, whether they are risk aversive, overall environmental concerns, and the sources of information that they encounter or listen to” (Burger and Gochfeld, 2006). Site (or region)-specific information on the reasons for fishing would allow for a communication strategy aimed at the local fishing population. Instead of just the consumption advisory information being provided to the public, risk managers must address multiple attitudes, behavioral patterns, and exposure pathways. Effective risk communication results in the target audience being provided with sufficient site- and fish- 26 specific information. This should include the risks and benefits o f consuming a given species of fish, at a given size or weight, so that they can make an informed decision (Burger and Gochfeld, 2006). Incorporating relevant information in enough detail to fully relay the message, but not so much detail that it is overwhelming and difficult to understand can be quite challenging. Groth (2010) developed a chart to organize the 51 seafood varieties into six groups based on Hg levels to serve as a framework for improving risk communication. He used FDA data on the Hg content o f each variety, and marketshare data from the National Marine Fisheries Service (NMFS), adapted by the FDA, to estimate contributions to the total amount o f Hg in the US seafood supply. He multiplied the mean Hg level by the share o f the market for each variety o f fish and shellfish to generate Hg input factors. These were indicative o f the relative inputs by each variety to the total amount of Hg in the US seafood supply. He then ranked the 51 seafood varieties by their relative contributions to total Hg, then sorted them into six categories by Hg content and examined risk communication implications o f the information thus generated. According to this chart, canned light tuna is categorized as “above average” for Hg content; however, the FDA’s advisory categorizes it as “low” for Hg content and recommends that pregnant women consume up to 12 grams per week (Groth, 2010). Groth (2010) identifies many deficiencies in current risk communication by the government: “ it does not address the needs o f consumers who eat a great deal o f fish; it offers no advice about numerous moderately high Hg fish that are significant sources of exposure if eaten regularly; it inaccurately describes the largest source o f Hg exposure in the American diet 27 as a “ low-Hg” fish; and it fails to draw distinctions among Hg levels in fish and shellfish varieties that strongly influence exposures for many consumers” (Groth, 2010). Many other studies attempt to reveal possible reasons for why or why not advisories are effective, and what can be done to make them more effective. Studies show that advisories are often ineffective at reaching ethnic groups, as well as fishermen with low income and educational levels (Tan et al., 2011). When advisories incorporate dissimilar priorities, it can increase the complexity o f the advice and send conflicting messages to the public. Tan et al. (2011) evaluated approaches o f consumption advisories to improve the effectiveness o f California advisories. They made several recommendations as a result o f their analysis, giving policy makers a few points to think about when creating an advisory. This research found that attempts to define portion size in quantities that depart from commonly consumed quantities to control fish intake are unlikely to be heeded; instead, advisories should place emphasis on the frequency o f consumption rather than portion size. They found that informants were more receptive to fish consumption advice when it was accompanied with information specific to the fish they were catching, particularly a visual depiction o f the fish’s Hg level; therefore, they recommend giving not just consumption advice, but Hg information for fish as well. They also recommend avoiding certain terms, symbols, and concepts that may cause confusion, as they found that for some informants, the inclusion o f one or more confusing terms was sufficient reason to disregard the entire material. Lastly, they recommend using portion sizes, Hg meters to convey contaminant levels, advice categories, and population definitions effectively. They found it was helpful to use Hg meter and portion size illustrations, and to group fish into three categories for high, moderate, or low Hg 28 levels. To ensure that the advisory will be effective on its target population, it is important to test the advisory materials among intended audiences before they are finalized (Tan et al., 2011). A North Carolina study assessed the determinants of subsistence fishing and tried to promote informed fish consumption among culturally distinct and lower income subsistence fishers (Driscoll et al., 2011). The study participants included African American, Hispanic, and Native American communities. Fish advisories were developed for each community to promote informed fish consumption intentions among residents who consume local fish and were successful in increasing knowledge and healthy intentions among most residents. The fish advisories were tri-fold brochures that were based on formative data collected in each community. Information that the brochures included was: a description o f the health and cultural benefits o f fishing and eating local fish; a description o f safe levels o f fish consumption for members o f various subpopulations; various methods for reducing exposure to MeHg without precluding local fish consumption completely; and contact information for local organizations and resources to which residents can go for more information. Further, the brochures included community-specific social values attributed to subsistence fish consumption, commonly held beliefs, and culturally sensitive mitigation strategies. All brochures were written at a sixth grade reading level using the Flesch-Kincaid readability program. The fish advisories were effective in educating those unaware o f the risk o f the existence o f MeHg in local fish. They also educated those who were aware o f the risk that popular measures intended to reduce exposure to the contaminant were ineffective (such as beliefs that the river cleanses itself o f MeHg or that the contaminant can be seen or removed in 29 preparation). The intention o f the advisories was not for the populations to cease consumption of locally caught fish altogether, but rather to continue eating locally caught fish with lower levels of MeHg. Only one o f the three communities actually intended to abide by the recommendations; the other two indicated that they intended to continue consuming locally caught fish without altering their consumption patterns. Burger and Gochfeld (2008) provided suggestions for future communication that “might improve the knowledge base for making decisions about fish consumption: 1) clearer statements about the agents causing the risk or benefit (e.g. Hg, omega-3 fatty acids), and the potential health outcomes (neurobehavioral deficits, lower cholesterol), 2) clearer statements about which fish are freshwater or saltwater fish (terms often used in advisories, but which are not generally understood), 3) clearer listing of which fish have high or low levels o f contaminants, specific to geographical region, and 4) target information to minorities about the factors contributing to risks and benefits, and about fish that are high or low in contaminants” (Burger and Gochfeld, 2008). 2.8 Risk assessment by FAO/WHO The FAO/WHO document titled “Food Safety Risk Analysis: A guide for national food safety authorities” was last updated in 2006. This document describes the structured Risk Analysis decision-making process with three distinct but closely connected components: risk assessment, risk management, and risk communication (Figure 2.2). 30 Figure 2.2 Generic components o f risk analysis (FAO/WHO, 2006) Communication Risk Management Scientific inputs Decisions involving policy and values Definitions for the three main components of risk analysis have been provided by FAO. Risk assessment: a scientifically based process consisting o f the following steps: i) hazard identification; ii) hazard characterization; iii) exposure assessment; and iv) risk characterization. Risk management: the process, distinct from risk assessment, o f weighing policy alternatives in consultation with all interested parties, considering risk assessment and other factors relevant for the health protection o f consumers and for the promotion o f fair trade practices, and, if needed, selecting appropriate prevention and control options. Risk communication', the interactive exchange o f information and opinions throughout the risk analysis process concerning risk, risk-related factors and risk perceptions, among risk assessors, risk managers, consumers, industry, the academic 31 community and other interested parties, including the explanation of risk assessment findings and the basis o f risk management decisions (FAO/WHO, 2006). Risk assessment and risk management are grounded in a science-based approach; risk assessment is the scientific component o f risk analysis, while risk management combines the scientific approach with other factors such as economic, social, cultural and ethical considerations (FAO/WHO, 2006). However, it is important for risk assessment to also involve judgments and choices that are not completely scientific, and for risk managers to clearly understand the scientific approaches used by risk assessors (FAO/WHO, 2006). 2.9 Risk management by FAO/WHO Risk management is best accomplished by using a systematic, consistent, and readily-understood framework while employing scientific knowledge on risk and other factors relevant to public health protection (FAO/WHO, 2006). The Food Safety and Risk Analysis document presents a generic risk management framework, which provides a practical, structured process for food safety regulators to apply the components o f risk analysis. There are three perspectives on risk that need to be addressed: technical, psychological, and sociological (FAO/WHO, 2006). The technological perspective is limited to scientific evaluation o f the likelihood and severity of harm. This may include an economic subset in which harm can be described in terms of health indices (FAO/WHO, 2006). The psychological perspective focuses on risk as a function o f individual perception. This takes various attributes into consideration such as willful exposure, ability to control risk, and catastrophic nature of risk, etc. (FAO/WHO, 2006). The sociological perspective views risk as a social and 32 cultural construct. The goal o f this perspective is to distribute costs and benefits in socially acceptable and equitable ways (FAO/WHO, 2006). The generic framework for risk management was designed to be functional in both strategic, long term situations and in the shorter term work o f food safety authorities. The framework is broken down into four parts which are all interconnected: preliminary risk management activities, identification and selection o f risk management options, implementation o f risk management decision, and monitoring and review (FAO/WHO, 2006). Please refer to Figure 2.3 for detailed descriptions of these four categories. Figure 2.3 Generic framework fo r risk management (FAO/WHO, 2006). Preliminary risk m anagem ent activities identify food safety issue develop risk profile establish goals of risk management decide on need for risk assessm ent establish risk assessm ent policy commission risk assessm ent, if necessary Identification and selection of risk m anagem ent options Monitoring and review monitor outcomes of controls) review controls) where hdicated identify possible options • evaluate options select preferred option(s) Implementation of nsk m anagem ent decision validate control(s) where necessary implement selected controls) verify implementation 33 2.10 Risk assessment tool for the Northern Health Authority The risk assessment tool developed in this thesis is designed to be unique to selected areas in the Northern Health Authority’s jurisdiction by providing local levels o f Hg in different fish species, local information on risk factors for elevated blood Hg, and a framework for information that should be collected on local patterns of fish consumption. This tool identifies which species from specific water bodies contain elevated Hg concentrations and how fish harvesting methods for these species can be adapted to minimize Hg exposure. The risk assessment tool is a combination of existing documents including the BC HealthFile # 68m (HealthLinkBC, 2011). This HealthFile was released to promote lowrisk fish consumption and to warn about fish consumption that may put British Columbians at risk for adverse Hg effects. The British Columbia Centre for Disease Control and Ministry o f Health issued guidelines specific to the province on the consumption of fish and Hg because it has fish consumption patterns that are unique in Canada, as there are many coastal, Aboriginal, and Asian communities in BC who tend to eat large amounts o f fish. Further, there is regional variation in Hg levels in fish available to consumers across Canada. Providing fish consumption advice can be a controversial issue, as it is undeniable that MeHg is toxic, but also undeniable that fish has many dietary benefits. Therefore, it is important to inform the residents o f Northern British Columbia about the types o f fresh fish that are available in the area and the risks and benefits associated with them. 34 3.0 Methodology 3.1 Overall approach 3.1.1 Source o f raw data. Data were collected from the British Columbia Ministry of Environment and Health Canada, which have offices located in Prince George, British Columbia. This included raw data that these agencies had collected, but not analyzed, as well as data obtained from reports that have already been published, or written for the sole use of the specific agency. In addition to Hg concentrations, fish type, fish length (tip of the snout to the tip of the longer lobe o f the fin), weight, sample location, and sample year were extracted from these data sets and used for this study. In some cases, data were duplicated between the various sources. To eliminate duplicated data, data sets were examined manually after they were entered into SPSS; redundant data sets were eliminated from the statistical analysis. 3.1.2 Use o f muscle tissue data only. Hg concentrations can vary between muscle tissue and organs. Most studies examining the health effects o f high Hg levels in fish on human health tend to use muscle tissue data, as it is the most widely consumed part o f the fish. Most of the data collected were from muscle tissue; however, there was a small amount o f data on Hg levels in organ, roe, water, and sediment that were omitted for consistency. 3.1.3 Use o f existing risk assessment tools. There are various risk assessment books and tools which were reviewed when creating the risk assessment tool for Northern Health. Existing government recommendations are incorporated into the tool and Health Canada’s guidelines for Hg levels and advised amounts o f consumption are used. The risk assessment framework that was adapted for this study is described in detail below. 35 3.2 Description o f study area Northern British Columbia is a vast area that consists o f many bodies o f water. For this study, 3097 fish samples were analyzed from 34 distinct areas (refer to Table 3.1 below). Currently in Northern British Columbia, there are only fish advisories for bull trout and dolly varden from Williston Lake and lake trout from Pinchi Lake (Environment Canada, 2010). 3.3 Fish collection and Hg analysis Fish samples were collected over a 26 year time span by different research groups; a variety o f methods were used to determine Hg concentrations in these samples. The data report total Hg concentrations; however, since total Hg in fish is comprised almost entirely of MeHg (Rasmussen et al., 2007), MeHg is used interchangeably with Hg in this study. All concentrations are presented on wet weight basis. Data on various fish species varied from location to location, and from year to year. The number o f fish samples collected from the water bodies also varied from year to year. 36 Table 3.1 Number offish collected from different w ater bodies used in this study Lake N B ab in e L ak e B ear L ak e 301 102 5 95 57 19 B row n L ak e C hu ch i L ak e C u n n in g h am L ake F ran co is L ak e G rassh am L ak e Inzana L ak e K azchek L ak e 28 65 46 ICemess L ak e M cK n ig h t L ak e N ations L ak e s N echako R e serv o ir - T a h tsa R e a c h N eco slie R iv e r - S tu art L ak e Pinchi L ak e P urvis L ak e Q uesnel L ak e R ain b o w C reek S tuart L ak e T akla L ake Tat chi R iv e r T chentlo T ezzero n L ake T o ch ch a L ak e T rem b leu r L a k e /M id d le R iv e r T sayata L ak e W eisn er L ak e W h itefish L ak e W iiliston R e se rv o ir - F in lay W illisto n R e se rv o ir - In g e n ik a W iiliston R e se rv o ir - W illisto n L ak e W illiston R e se rv o ir - P arsn ip 35 8 8 16 26 162 20 110 11 141 34 72 62 191 11 105 191 40 66 334 W illiston R e se rv o ir - P e ac e 180 42 166 304 W itch L ak e Total 44 3097 3.4 Statistical analysis o f data The program SPSS for Windows, version 16 (SPSS, Chicago, Illinois, USA) was used for data analysis. Before analysing the data, several aspects o f Hg fish tissue concentrations were o f interest. For example, did some water bodies exhibit higher concentrations o f Hg in fish tissue, on average, than fish caught in other water bodies? 37 Did Hg concentrations of specific fish species exhibit differences between water bodies? How did these relationships change over the study period (1974 —2000)? The data set had several major limitations with regards to suitability for the tests used to measure these differences, and will be clarified below. The various sources o f data had information on the fish species, length, weight, Hg concentration, location and year collected. In some cases the specific fish type was given whereas in others only the general fish type is provided (for example, Sockeye Salmon in a specific case and Salmon in a general). In these cases, the fish types were not combined so that if there was significance in a specific type it could be identified. Also, some of the data did not report fish length (only 1801 fish had length values in a data set of 3097 samples). Temporal and spatial trends, as well as species, weight, and length differences were all assessed in comparison to the Hg concentrations. 3.4.1 Controlling for length:weight relationships As mentioned above, less than half of the fish data (41%) reported fish length, while all data sources reported fish weight. In these samples it was debated whether the lengths should be estimated for the data analysis since this is the main method o f identifying size of fish in many other types o f studies. Length data that were available were correlated with weight and a linear equation was formed (y = mx + b). Since weight was available for all fish, it was used in the equation to estimate length values for all those without lengths already. There are obvious limitations with this technique and it was decided that it is much more accurate to use the actual weights of the fish instead o f the estimated lengths. 38 3.4.2 ANOVA/ANCOVA In order to carry out ANOVA and ANCOVA tests, data must meet certain criteria. For example, dependent variables should be normally distributed. Neither Hg concentration nor the weight o f fish met this criterion. Both variables had extremely large values for skewness and kurtosis (see Table 10.7 in Appendix) which were also confirmed by statistical tests for normality (Kolmogorov-Smimov). A square root transformation was applied to the weight, and a cube root transformation applied to Hg concentration (Tables 10.5 and 10.7). However, even having applied these transformations, the Kolmogorov-Smimov test showed that they were still non-normal (Tables 10.6 and 10.8). Another assumption is that each variable should have a fairly similar variance. This was not the case as shown by Levene’s test. For ANCOVA, there are several other assumptions in addition to those o f the ANOVA. Covariates must be linearly related to the dependent variable; this assumption was met as weight and Hg concentration were positively correlated. An assumption that was violated for this data set was that the covariate should be unrelated to the independent variable, in this case the location. Weight was highly correlated with location, which can be explained by the fact that larger bodies o f water are likely to support larger fish, whereas smaller bodies o f water are not able to support larger fish. Also the growth rate and trophic structure can vary among lakes due to their specific geochemistry. Another assumption is that covariates must have a homogeneity o f regression effect. Essentially, there have to be equal effects on the dependent variable across all different independent variable subgroups (all slopes have to be equal). In a scatterplot of weight versus Hg concentration with location as the control variable (data 39 not shown), it was clear that slopes were extremely different, and thus it was decided that ANCOVA should not be carried out. Given the fact that the data violates almost all o f the assumptions for ANOVA/ANCOVA, doing either o f these analyses will provide results that would be difficult to interpret, and worse, misleading and/or inaccurate. Consequently, ANOVA and ANCOVA were not conducted in this thesis. Instead, descriptive statistics were used and in some cases correlations were conducted, when appropriate. 3.4.3 Use o f data in the Risk Assessment Tool The results that were depicted by the data analysis were used to develop consumption advice for specific areas. The Hg concentration in fish was displayed by year, water body, species, and size o f fish. This was used to identify if Hg concentrations have declined and whether it is now safe to consume fish from specific water bodies, whether the same species have elevated Hg levels in various water bodies, and whether fish size made a significant difference in Hg concentration. Answering these questions allowed us to suggest whether the public can lower Hg intake by eating different species of fish from certain lakes or by fishing for their desired species in a different lake (if Hg levels are high for that species in their usual fishing lake). To develop the risk assessment tool, the mean Hg concentration o f all species (with an N=60 or greater), was adjusted to standardize weight for each species. This was done so that the variability o f Hg concentration would only reflect the difference in location. Using the results of this, the risk assessment tool is able to predict which species should be avoided from which lakes. 40 4.0 Results 4.1 Fish species and water bodies included in this study A total o f 20 types of fish were sampled from 34 water bodies (Table 3.1), ranging from a sample size o f 2 (sturgeon) to 892 (lake whitefish). The total number o f fish samples used in this study was 3097. O f these, 1801 had both length (mm) and weight (kg) values; the remainder had only weight. Samples were collected between 1974 and 2000 (samples were collected in 16 o f these 26 years). 4.2 Fish weights and fish lengths Figure 4.2.1 displays the number o f fish caught as well the mean weight o f those fish for the year. The data presented highlights several important points. There were a number of years in which very few fish were caught (1975, 1985, 1989, 1992, 1993, 1996), the average weight of the fish caught was very low (less than 0.742 kg). The only exception to this rule appears to be 1986, in which only 58 fish were caught, but had a mean weight of almost 2 kg (greatest mean weight in this study). The trends in mean fish weight versus the location at which the fish were caught are shown in Figure 4.2.2. Fish caught in Tochcha Lake (lake trout in 1986) had the highest mean weight at almost 3.4 kg. However, only 11 fish were caught at this location. Quesnel Lake was next with a mean weight o f 2.9 kg, with a total o f 110 fish caught (rainbow trout and lake trout in 1988). Brown lake (dolly varden) and Rainbow Creek (rainbow trout) had the lowest mean weights, with 0.11 and 0.05 kg respectively. Total fish caught in these locations were 5 and 11, respectively. 41 Figure 4.2.1 Mean weight (kg) o f fish caught vs. year caught (n=3097) 2.5 700 7 4. 600 1.5 ■M 400 0J5 '3 1 300 I■2 c ~ n Total Number of Fish Caught i z 200 0.5 100 0 I I I I 1 i I 1 I i 42 ,Mean W eight in kilograms Figure 4.2.2 Mean weight (kg) o f fish caught vs. location (n=3097) 350 3.5 300 250 2.5 200 .5Cl? 5 150 1.5 Q M ean W eight in kilograms B T o ta J N u m b e ro f Fish Caught 100 1 -- 0.5 - • s s s s s y s s jy s s s y fS o / / / / Y Jit* Location 43 s Figure 4.2.3 shows the mean weight of each fish species caught over the study period. Lake whitefish were by far the most common fish caught at 892 samples, followed by lake trout (n=660), and rainbow trout (n=365). Many species were caught for which sample sizes are extremely small: coho salmon (n=4), large scale sucker (n=3), rocky mountain whitefish (n=l), squawfish (n=l), peamouth chub (n=3) and sturgeon (n=2). The mean length of fish caught by year is displayed in Figure 4.2.4. It can be seen that the highest mean length o f fish caught was in 1978 at 518 mm (n=38), followed by 1981 at 505 mm (n=101). Lowest average lengths were seen in 1989 (93 mm, n = l) and 1992 (197 mm, n=35). Note: missing lengths were not calculated using the calculation noted in methods section for this chart; total n = 1801. Figure 4.2.5 illustrates the mean length o f fish by the location in which they were caught. The highest mean lengths were caught in Quesnel Lake (628 mm, n=55) and Witch Lake (613 mm, n=3). Lowest mean lengths were seen in Kemess Lake (197 mm, n=35), Brown Lake (217 mm, n=5), and Purvis Lake (317 mm, n=3). One fish (a Rainbow Trout) was caught in Rainbow Creek with a length o f 93 mm. It can be seen in Figure 4.2.6 that lake trout (581 mm, n=370), salmon (595 mm, n=13), sockeye salmon (560 mm, n=69), and char (562 mm, n=47) had the highest mean length. Fish with the lowest mean length were peamouth chub (253 mm, n=3) and mountain whitefish (260 mm, n=33). 44 Figure 4.2.3 Mean weight (kg) o f fish caught vs. fish type (n=1801) ■ M ean W eight in kilograms ■ Total N um ber o f Fish Caught Fish Type 45 Figure 4.2.4 Mean length (mm) o f fish caught vs. year caught (n=1801) 4S0 600 400 soo - 350 300 400 C— a Total N um ber of Fish C aught - 200 ISO - 100 200 - 1974 1976 1977 1978 1979 1980 1981 1989 1986 Year 46 1992 1996 2000 100 * ■ ' M ean Length In m m Figure 4.2.5 Mean length (mm) o f fish caught vs. location caught (n—1801) 250 700 600 - 200 500 150 82 400 e 1 E 3 300 ~ 100 ■ M ean Le ngth of Fish in m m ■ Total N um ber of Fish Caught 200 - 100 N?'Jp ^>*P / 'iP »P / /✓ g» hP .AO .'iK >4p \?hP V>iP ? / nP vfsP ''iP >*P N\ip ? / sP NsP? e?sP * / / j? ^ y S /s s s /y Location 47 * y /jy // Figure 4.2 .6 Mean length (mm) o f fish caught vs. fish type (n-1801) 700 600 500 400 c 400 - 300 ■ M ean Lenth in m m 200 ■ T o ta lN u m b e r o fF ish C aught 100 r J- a '- ' > ^ V> (f 0° o. ^9 ^ ^ iV .V * X>° •H- Fish Type 48 4.3 Mean Hg concentrations in fish by year, location, and species o f fish Mean Hg concentrations in this data set varied from year to year, by location, and also by species o f fish. The figures below illustrate the differences. As can be seen in Figure 4.3.1, fish caught in 1974 had the highest mean Hg concentration at 2.11 ppm (n=42). The second highest year for Hg levels was 1986, with a mean Hg concentration of 0.84 ppm (n=58). In years such as 1979 and 1988 where sample sizes were large (n=675 and n=750, respectively), Hg levels were relatively low at 0.24 ppm and 0.30 ppm, respectively. Mean Hg concentration were found to be the lowest in 1992 (0.03 ppm, n=35), 1989 (0.09 ppm, n=l 1), and 1993 (0.09 ppm, n=16). Figure 4.3.1 Mean Hg concentration (ppm) o f fish caught v.v. year caught (n=3097) ------- --------------------- ------- -------------------------- r 300 2.5 700 2 600 500 1.5 E z s a Total N u m b e r o f Fish C au gh t 1 - 300 - 200 0.5 100 0 0 ■ ^ • t - n o f ^ o Q c r t O t - i c n c o o o c n c s f r o t o o rCs F» rf ^Cf i^ cr *r ^t oo ri^corojocoioooco ioco fo tcoo co nocor > c cn ro»oo rH tH t*H t—1 rH rH «—1 rH tH H rH rH pH 49 cH t-H fNJ ~ Mean Mercury Concentration in ppm Figure 4.3.2 presents mean Hg levels by location and clearly displays that fish from Pinchi Lake have the highest mean Hg concentration (1.17 ppm, n=162). All o f the fish species caught from Pinchi Lake were at or above the Health Canada’s Hg reference dose of 0.5 ppm, with the exception o f rainbow trout, which had a mean Hg concentration o f 0.36 ppm. McKnight Lake also has a very high Hg concentration at 0.53 ppm; however, the sample size is very small with only eight fish caught at that location. Areas with lower mean Hg concentrations are Kemess Lake ( 0.03 ppm, n=35), Nechako River - Tahtsa Reach (0.09 ppm, n=16), and Inzana Lake (0.13 ppm, n=65). The average Hg concentration for all fish caught over the study period was 0.30 ppm (n=3097). Large scale sucker and peamouth chub (both caught in Pinchi Lake) had the highest mean Hg levels at 2.59 ppm and 1.90 ppm respectively (Figure 4.3.3); however, it can also be seen that they have a very small sample size (n=3 for both). Bull trout had mean Hg concentration higher than 0.5 ppm at 0.70 ppm (n=313). However, almost all o f the bull trout samples were taken from the Williston Reservoir from various reaches (Finlay had the biggest sample size o f 160). The lowest concentration of mean Hg was found in coho salmon (0.04 ppm, n=4), salmon (0.07 ppm, n=13), kokanee (0.085 ppm, n=195), and mountain whitefish (0.10 ppm, n=60). 50 Figure 4.3.2 Mean Hg concentration (ppm) o f fish caught vs. location caught ® Mean Mercury C oncentration in ppm ///>\ vv%vv%vvvvv^^^^ 4 Location 51 NJ Ln Ln Bull T ro u t B urbot Char Coho Salm on Dolly V arden Kokanee Lake T rout Lake W hitefish Large Scale Sudcer H zr < T3 re M o u n ta in W h itefish |S" P e a m o u th Chub R ainbow T rout 0 Rocky Mt. W h itefish S alm on S ockeye Salm on K> Squaw fish S tu i^ e o n Sucker Trout W hite S uck er W hitefish O m W CO ^3. Ln cn •si eo ui t-* O O O S 0 «O © O D3 ft O Q O O fi O O 0 £2 N u m b e r o f Fish "H Q n? aD> C CF nOI c c n Q 3 n rp 3 r-h a. o3 T3 T3 3 Figure 4.3.3 Mean Hg concentration (ppm) of fish caught vs. type of fish (ppm] p 4.4 Relationship between Hg concentration andfish size The largest sample size of lake whitefish was from the Peace Reach o f Williston River (N=158, 0.130 ppm); however, the heaviest lake whitefish were found in W eisner Lake (N=20) at 1.09 kg, with a Hg concentration o f 0.086 ppm. The smallest lake whitefish were caught in the Parsnip Reach o f Williston River (0.262 kg) with a mean Hg concentration of 0.190 ppm. The highest levels o f Hg in lake whitefish were found in Pinchi Lake, with a mean concentration o f 0.495 ppm (0.535 kg). The heaviest lake trout were found in Whitefish Lake (N=40) at a mean o f 2.99 kg, however, the Hg concentration was only 0.311 ppm whereas the lake trout from Pinchi Lake (N=75) had a lower weight of 2.74 kg, but a Hg concentration o f 1.82 ppm. The smallest lake trout were caught in Purvis Lake (N=6), at 0.35 kg and a Hg concentration of 0.308 ppm. The smallest rainbow trout were caught in Rainbow Creek (N=l 1) at 0.0512 kg (0.0856 ppm Hg). The heaviest rainbow trout were 3.02 kg from Quesnel Lake (N=73), with a Hg concentration of 0.117 ppm. However, the rainbow trout with the highest Hg concentrations were also from Pinchi Lake (N=13) at 0.36 ppm and a weight o f 0.309 kg. Figure 4.4.1 demonstrates that the heavier the fish, the higher the mean Hg concentration. However, the lower and mid-weight categories do have a larger sample size; fish over 3 kg only account for approximately 11 % o f the total data set. 53 [p p m ] Figure 4.4.1 Mean Hg concentration (ppm) in fish tissue vs. fresh weight (gms) offish 0 fit ' 0.193 0.200 ° ’237 II I I /^ ^/ / / ^/ ^/ ^/ / n f i f f i f f i f f ' rfi ^

0.5ppm in bull trout from Williston Reservoir in the year 2000 (MWLAP, 2002). The fish species that were above the Health Canada guideline of 0.5 ppm for Hg concentration in W illiston Lake were dolly varden and bull trout, the same two species that have an ongoing consumption advisory. 7.2 Hg levels in fish tissue varied with fish size and sample period Besides location and trophic level, there are many other factors that can influence Hg concentrations in fish. Fish size is one o f the well-known influencing factors for elevated Hg levels. The heaviest species was char, followed by lake trout, salmon, sockeye salmon, and bull trout. The four longest species were the same, although in a different order: salmon, lake trout, sockeye salmon and char. Fish length did not seem to be as closely related to weight as would be expected. When comparing years o f data collection, the heaviest fish were caught in 1986 and in between 1976 and 1978, whereas the longest fish were caught between 1978 and 1981. An explanation for this may include fish growth rate, which is the temporal change in either fish weight or length. Fish growth 65 rate has been shown to influence Hg accumulation (and in turn, Hg biodilution) in fish muscle, as faster-growing fish have been shown to have lower Hg concentrations than slower-growing fish at a certain length (Lavigne et al., 2012). Biodilution is defined as a reduced overall accumulation o f a contaminant within an organism due to an increase in body size resulting from differences in bioenergetic processes. Hg biodilution has been shown to partly explain decreased Hg concentrations in fish when it was not explained by changes in fish diet, structural alterations o f the trophic web, a reduction o f MeHg levels in forage fish, or by a reduction in whole-lake MeHg content (Lavigne et al., 2012). Fish growth rate has many influencing factors including the ratio between primary watershed area and lake area, the ratio between drainage area and lake area, riparian wetland coverage, land use and vegetation coverage o f the primary watershed, water quality variables and the sportsfishing intensity. Lavigne et al. (2012) found that growth rate could be used as an integrated proxy to predict Hg concentration in fish muscle in two slower-growing species (walleyes and northern pike) which had higher Hg concentrations at standardized length. Thus, they concluded that proper control o f fish growth rate through fishing pressure, lake ecology, and watershed management could be used to minimize the toxic risk associated with Hg exposure from fish consumption. Quesnel Lake had the longest fish and heaviest fish (mean weight o f 2.87 kg); however, Quesnel Lake did not have the highest mean Hg concentration (0.16 ppm, N =110). In comparison, Pinchi Lake had the highest mean Hg concentration (1.17 ppm, N= 162) and a mean weight o f 1.51 kg. Only lake trout (N= 73) and rainbow trout (N=37) were sampled from Quesnel Lake, whereas 8 different species were sampled from Pinchi Lake, the majority of which were lake trout (N=75) and lake whitefish 66 (N=51). The lake trout from both lakes had similar sample sizes and similar weight, however the Hg concentrations for this species differed greatly in the water bodies, with Pinchi Lake’s lake trout having a mean Hg concentration of 1.82 ppm, and Quesnel Lake’s having only 0.24 ppm. It is demonstrated in Figure 4.4.1 that the mean Hg concentration in the fish in this study (n=3097) is moderately positively correlated with the weight o f fish (r - 0.316, p < 0.0001). In fish over 2 kg, the Hg concentration is very close to or over Health Canada’s maximum of 0.5 ppm guideline. This weight category includes char, lake trout, salmon, and sockeye salmon. This finding is supported by many previous studies, including a study completed by Storelli in 2007 which analyzed Hg concentration in fish versus their size. It was found that there was a significant relationship between Hg concentration and fish size for all species (Storelli, 2007). However, it is noted by Bhavsar (2010) that Hg concentrations in fish typically increase with age, and that fish size is obtained as a surrogate measure for the duration o f contaminant exposure because it is easy and inexpensive to acquire (Bhavsar, 2010). 7.3 Hg levels in fish tissue varied with species Another possible contributing factor to varying Hg concentrations in fish is species. Figure 4.4.3 illustrates mean Hg concentration by fish species. Unfortunately, the two species (large scale sucker and peamouth chub) with the highest concentrations both have very small sample sizes o f only 3 fish each. Bull trout (n=313) had a concentration of 0.70 ppm and dolly varden (n=87) o f 0.50 ppm. It is interesting to note that bull trout and dolly varden were considered to be the same species until 1980, when they were reclassified as a separate species (U.S. Fish and Wildlife Service, 1998). Piscivorous fish 67 (such as lake trout and bull trout) tend to have higher concentrations of Hg than fish that consume plankton (such as kokanee and lake whitefish) (Baker, 2002), which is supported by the results o f this study. It was found that the highest Hg concentrations in fish in this study were in those caught in 1974. Upon closer examination, as displayed in Table 10.15 in the appendix, all o f the fish sampled in 1974 came from Pinchi Lake. Further, the second highest Hg levels were in 1986, which was also exclusively from Pinchi Lake. Pinchi Lake was not part o f any of the sampling areas in those years in which Hg levels were the lowest. 7.4 Limitations o f this study Although we are able to draw conclusions from this study, there are many limitations which prevent us from making any strong statements. This is mainly attributed to lengthy time span over which the data were collected; this can be seen as a positive, as it gives us an idea whether fish Hg concentrations are decreasing over time. However, due to the lack of consistency in location, this is a hard pattern to accurately conclude. For instance, when looking at Pinchi Lake, it is evident that Hg levels remained high between 1974 and 2000; however, samples were only taken in four o f those years. The last two years the fish were sampled in (1986 and 2000) had a 14 year gap and they had sample sizes of only 47 and 67. In order to truly capture the patterns o f mean Hg concentrations in fish over the years, it would be important to take a specific number of samples closer together and from the same location(s). Another large factor that limited the analysis was that almost half o f the data set was missing fish length. As discussed earlier, due to the variability of fish weight compared to fish length, length is the most reliable method o f accurately estimating fish 68 size. Since we are correlating fish size with mean Hg concentration, this may contribute to an incorrect analysis. That said, lengths and weights o f fish were highly correlated; therefore, using the weight is not completely inappropriate. Further, the data were collected from many different agencies and existing papers; therefore, the sampling techniques may have varied quite widely. This would include the actual lab procedures used to determine the Hg concentrations, as well as the method used to weigh and measure the fish. 7.5 Development and potential application o f the Risk Assessment Tool By using the framework developed by FAOAVHO for risk assessment and management, it was clear that a risk assessment tool for the MeHg levels in fish in Northern BC would be o f benefit to those who fish in the local waters and for the public health officials that provide advice and guidance. However, due to the limitations o f the data set, development o f the risk assessment tool presented some difficulties. By adjusting for weight (standardizing the data for differences in fish weight), the data could be presented so that each sample could be viewed as either above or below the cut-off for excessive Hg concentration. A few tables were developed which can be used as a reference by Public Health professionals and the general public. Table 6.2 would be very useful to those trying to decide whether to consume a certain species from a certain lake, as each species is listed separately with the percentage of contaminated samples from that lake. This is important because it does not limit all fish consumption from that water body, but allows one to determine which species is a better choice. Further, the number o f samples that the percentage is derived from is listed as well, which assists in making an informed 69 decision; for instance, if only 3 samples were taken for that species from that lake and 100% were contaminated, farther investigation may be warranted. However, if 100 samples were taken and 90% were contaminated, then it is pretty certain that caution should be used when consuming that species taken from that lake. Further, to simplify for fishers that are not interested in the percentages, Table 6.3 includes a list of fish species that are considered most contaminated from specified water bodies and from which nearby water bodies contain less contaminated fish. This does not limit consumption o f a specific species, but instead provides safer options; also, if it is a specific lake that is the desired fishing spot, then the fisher is aware of species that should not be consumed. 8.0 Conclusion Several interesting and important conclusions can be drawn from this study. Studying Hg concentrations in fish over such a large time span and in many different water bodies, allows us to identify concerns for high fish Hg concentrations. This study shows that as of the year 2000 Hg concentrations in fish were still high in some areas o f Northern British Columbia. In Canada, there have been reductions in Hg emissions in base metal mining, Hg used in the manufacture o f chlorine and pesticides has mostly been eliminated, and releases from paints and batteries have also declined. Further, the Canadian Council of Ministers o f the Environment (CCME) has identified Hg as a priority issue (MWLAP, 2002 ). 70 There has been a considerable amount of information collected on fish Hg concentrations in Northern British Columbia since 1970; however, this information has been widely dispersed amongst many government and private agencies. There are other summaries o f fish Hg concentrations in Northern British Columbia similar to this one; however, this study has some data that is not included in those others. Further, this study did not eliminate data points based on size, or missing lengths. They were accounted for and noted in the analysis, but it was believed that eliminating certain data did not provide a clear picture of the actual results. Despite the large amount o f information available on fish Hg concentrations in this area, there is no ongoing monitoring program. It is highly recommended that a continuing and systemic monitoring program be put into place to update the existing and future fish consumption advisories accurately. Most importantly, this study has produced a risk assessment tool which allows public health officials and members of the public to be able to make informed decisions about which water bodies they are fishing from. When fishing for a specific species the public is able to refer to this tool and choose to fish from a lake that has been found to have lower Hg concentrations in that specific species. 71 9.0 References Adriano, D.C. (2001). Trace Elements in Terrestrial Environments, 2nd Edition. Springer Verlag, New York, New York. ASTDR (Agency for Toxic Substances and Disease Registry) (2011). Detailed Data Table fo r the 2011 Priority List o f Hazardous Substances that will be the Subject o f Toxicological Profiles. Retrieved from http://www.atsdr.cdc.gov/SPL/resources/ATSDR 2SPL Detailed Data Table.pdf on September 11, 2012. Baker, D,, Young, J., & Arocena, J. M. (2000). An integrated approach to reservoir management: The Williston Reservoir case study. Environmental Management, 25(5), 565-578. Baker, R. (2002). Fish Mercury Database Summary - 2001. Prepared for B.C. Hydro by Aqualibrium Environmental Consulting Inc., Vancouver, BC. Barrett, J.R. (2010). An Uneven Path Forward: The History o f Methylmercury Toxicity Research. Environmental Health Perspectives, 118(8), 1137-1145. Beyrouty, P., & Chan, H. M. (2006). Co-consumption o f selenium and vitamin E altered the reproductive and developmental toxicity o f methylmercury in rats. Neuro toxicology and Teratology, 28(1), 49-58. Bhavsar, S.P., Gewurtz, S.B., McGoldrick, D.J., Keir, M.J., and Backus, S.M. (2010). Changes in Mercury Levels in Great Lakes Fish Between 1970s and 2007. Environmental Science and Technology, 44, 32 73-3279. Bhavsar, S.P., Awad, E., Mahon, C.G., Petro, S. (2011). Great Lakes fish consumption advisories: is mercury a concern? Ecotoxicology, 20, 1588-1598. Bodaly, R.A., Jansen, W.A., Majewski, A.R., Fudge, R.J.P., Strange, N.E., Derksen, A.J., and Green, D.J. (2007). Postimpoundment time course o f increased mercury concentrations in fish in hydroelectric reservoirs o f northern Manitoba, Canada. Archives o f Environmental Contamination and Toxicology, 53(3), 379—389. Burger, J., and Gochfeld, M. (2005). The Peconic River: Concerns Associated with Different Risk Evaluations for Fish Consumption. Journal o f Environmental Planning and Management, 48(6), 789-808. Burger, J., and Gochfeld, M. (2006). A framework and information needs for the management of risks from consumption o f self-caught fish. Environmental Research, 101, 275-285. 72 Burger, J., and Gochfeld, M. (2008). Knowledge about fish consumption advisories: A risk communication failure within a university population. Science o f the Total Environment, 390, 346-354. CBC News (2012). Mercury poisoning effects continue at Grassy Nanows. Posted on June 4, 2012. Retrieved from http://www.cbc.ca/news/technologv/storv/2012/06/04/grassv-narrowsmercurv.html on September 11, 2012. Chan, H. M., Scheuhammer, A. M., Ferran, A., Loupelle, C., Holloway, J., & Weech, S. (2003). Impacts o f Mercury on Freshwater Fish-Eating Wildlife and Humans. Human & Ecological Risk Assessment, 9(4), 867-883. Chan, L., Receveur, O., Sharp, D., Schwartz, H., Ing, A., and Tikhonov, C. (2011). First Nations Food, Nutrition and Environment Study (FNFNES): Results from British Columbia (2008/2009). Prince George: University o f Northern British Columbia. Clarkson, T.W. (2002). The Three Modem Faces of Mercury. Environmental Health Perspectives, 110 (S-l), 11-23. Cohen, J.T., Bellinger, D.C., Connor, W.E., Kris-Etherton, P.M., Lawrence, R.S., Savitz, D.A., et al. (2005). A quantitative risk-benefit analysis o f changes in population fish consumption. American Journal o f Preventative Medicine, 29, 325-334. Couzin, J. (2007). Dietary guidelines spark flap over fish consumption. Science, 318, 550-551. DeWeese, A.D., Kmiecik, N.E., Chiriboga, E.D., Foran, J.A. (2009). Efficacy o f RiskBased, Culturally Sensitive Ogaa (Walleye) Consumption Advice for Anishinaabe Tribal Members in the Great Lakes Region. Risk Analysis, 29(5), 729-742. Driscoll, D., Sorensen, A., Deerhake, M. (2012). A Multidisciplinary Approach to Promoting Healthy Subsistence Fish Consumption in Culturally Distinct Communities. Health Promotion Practice, 13(2), 245-251. EFSA (European Food Safety Authority) Journal (2010). Risk Assessment on M ercury in Fish. Retrieved from http://www.efsa.europa.eu/en/press/news/contam040318.htm on July 10, 2012. Environment Canada (2010). Freshwater Fishing Regulations Synopsis 2009-2011. Retrieved from http://www.env.gov.bc.ca/fw/fish/regulations/docs/0911/fishsvnopsis 2009-11 .pdf on September 11, 2012. FAO/WHO (Food and Agriculture Association o f United Nations/World Health Organization) (2006). Food Safety Risk Analysis: A guide fo r national fo o d safety 73 authorities. Retrieved from ftp://ftp.fao.org/docrep/fao/009/a0822e/a0822e0Q.pdf on September 11, 2012. FAO/WHO (Food and Agriculture Association o f United Nations/World Health Organization) (2011). Report of the Joint FAO/WHO Expert Consultation on the Risks and Benefits o f Fish Consumption. Rome, Food and Agriculture Organization o f the United Nations; Geneva, World Health Organization, 50pp. FAO Fisheries and Aquaculture Report No. 978. FNFNES (2012). First Nations Food, Nutrition and Environment Study. Prince George: University o f Northern British Columbia. Retrieved from http://www.fhfnes.ca/ on September 4, 2012. Fry, B. and Chumchal, M. (2012). Mercury bioaccumulation in estuarine food webs. Ecological Applications, 22(2), 606-623. Ginsberg, G.L., and Toal, B.F.(2008). Quantitative approach for incorporating methylmercury risks and omega-3 fatty acid benefits in developing speciesspecific fish consumption advice. Environmental Health Perspectives, 117(2), 267-275. Groth, E. (2010). Ranking the contributions o f commercial fish and shellfish varieties to mercury exposure in the United States: Implications for risk communication. Environmental Research, 110, 226-236. Harada, M., et al., 2011. Mercury Pollution in First Nations Groups in Ontario, Canada: 35 years o f Canadian Minamata Disease. Journal o f Minamata Studies, 3, 3-30. Health Canada (2007a). Human Health Risk Assessment in Mercury in Fish and Health Benefits o f Fish Consumption. Retrieved from http://hc-sc.gc.ca/fnan/pubs/mercur/merc fish poisson e.html on September 11, 2012. Health Canada (2007b). Health Canada’s revised assessment o f mercury in fish enhances protection while reflecting advice in Canada’s Food Guide. Retrieved from http://www.hc-sc.gc.ca/ahc-asc/media/advisories-avis/ 2007/2007 31-eng.php on September 11, 2012. Health Canada (2011). Mercury in Fish - Questions and Answers. Retrieved from http://www.hc-sc.gc.ca/fn-an/securit/chem-chim/environ/mercur/merc fish qapoisson qr-eng.php on September 11, 2012. HealthLinkBC (2011). Healthy Eating: Choose Fish Low in Mercury. Nutrition Series, No. 68m. Retrieved from http://www.healthlinkbc.ca/healthfiles/pdf/hfile68m.pd f on September 11, 2012. 74 Hobson K. (2006). Diet, Fitness, and Exercise: Two Reports Offer Guidance on Eating Fish. US News and World Report. Retrieved from http://health.usnews.com/usnews/health/articles/061019/19fish.htm on September 11 , 2012 . ILEC (International Lake Environment Committee) (2008). Williston Lake. Retrieved from http://www.ilec.or.jp/database/nam/nam-29.html on April 26, 2012. ILSI (International Life Sciences Institute) (2012). BRAFO. Retrieved from http://www.ilsi.org/europe/Pages/ViewItemDetails.aspx?ID:=20&ListName=Publi cations on July 10, 2012. Jacobson, J.L. (2001). Contending with contradictory data in a risk assessment context: the case o f methylmercury. Neurotoxicology, 22(5), 667-675. Karagas, M., et al. (2012). Evidence on the Human Health Effects of Low-Level Methylmercury Exposure. Environmental Health Perspectives, 120(6), a224a225. Kinghom, A., Solomon, P., and Chan, H.M. (2007). Temporal and Spatial Trends o f mercury in fish collected in the English-Wabigoon river system in Ontario, Canada. Science o f the Total Environment, 372(2-3), 615-623. Lando, A.M., and Zhang, Y. (2011). Awareness and knowledge o f methylmercury in fish in the United States. Environmental Research, 111, 442-450. Lavigne, M., Lucotte, M., and Paquet, S. (2010). Relationship between Mercury Concentration and Growth Rates for Walleyes, Northern Pike, and Lake Trout from Quebec Lakes. North American Journal o f Fisheries Management, 30(5), 1221-1237. Legrand, M., Arp, P., Ritchie, C., & Chan, H. M. (2005). Mercury exposure in two coastal communities o f the Bay o f Fundy, Canada. Environmental Research, 98(1), 14-21. Mahaffey, K.R., Sunderland, E.M., Chan, H.M., Choi, A.L., Grandjean, P., Marien, K., Oken, E., Sakamoto, M., Schoeny, R., Weihe, P., Yan, C., and Yasutake, A. (2011). Balancing the benefits o f n-3 polyunsaturated fatty acids and the risks o f methylmercury exposure from fish consumption. Nutrition Reviews, 69(9), 493508. Mast, M.A., and Krabbenhoft, D.P. (2010). Comparison of mercury in water, bottom sediment, and zooplankton in two Front Range reservoirs in Colorado, 2008-09. U.S. Geological Survey Scientific Investigations Report, 2010-5037. 15 Melwani, A.R., Bezalel, S.N., Hunt, J.A., Grenier, J.L., Ichikawa, G., Heim, W., Bonnema, A., Foe, C., Slotton, D. G., Davis J.A. (2009). Spatial trends and impairment assessment of mercury in sport fish in the Sacramento-San Joaquin Delta watershed. Environmental Pollution, 157, 3137-3149. MWLAP (Ministry o f Water, Land, and Air Protection) (2002). Environmental Trends in British Columbia, State o f Environment Reporting. Retrieved from http://www.env.gov.bc.ca/soe/et02/ET2002.pdfon October 21, 2012. Mousavi, A., Chavez, R.D., Ali, A.S., Cabannis, S.E. (2011). Mercury in Natural Waters: A Mini-Review. Environmental Forensics, 12, 14-18. NRC (National Research Council) 2000. Toxicological Effects o f Methyl mercury. National Academy Press, Washington DC. Oken E., Kleinman, K.P., Berland W.E., Simon, S.R., Rich-Edwards, J.W., and Gillman, M.W. (2003). Decline in Fish Consumption Among Pregnant Women After a National Mercury Advisory. The American College o f Obstetricians and Gynecologists, 102(2), 346-351. Perez, H., Sullivan, E.C., Michael, K., Harris, R. (2012). Fish Consumption and Advisory Awareness Among the Philadelphia Asian Community: A Pilot Study. Journal o f Environmental Health, 74(8), 24-28. Price, C., Goodson, B., and Stewart, G. (2007). Infant and Environmental Exposure to Thimerosal and Neuropsychological Outcomes at Ages 7 to 10 Years. Prepared for: Centers for Disease Control and Prevention, Technical Report: Volume 1. Rasmussen, R.S., Nettleton, J., Morrissey, M.T. (2005). A review of mercury in seafood: special focus on tuna. Journal o f Aquatic Food Product Technology, 14(4), 71100 . Rasmussen, P.W., Schrank, C.S., and Campfield, P.A. (2007). Temporal Trends in Wisconsin walleye (Sander vitreus), 1982-2005. Ecotoxicology, 16, 541-550. SACN/COT (Scientific Advisory Committee on Nutrition/Committee on Toxicity) (2004). Advice on fish consumption: benefits and risks. Retrieved from http://cot.food.gov.uk/cotreports/cotiointreps/sacnfishconsumption on September 11 , 2 0 1 2 . Sassi, F. (2006). Calculating QALYs, comparing QALY and DALY calculations. Health Policy and Planning, 21(5), 402-408. Shimshack, J.P., Ward, M.B. (2010). Mercury advisories and household health trade-offs. Journal o f Health Economics, 29, 674-685. 76 Stockner, J., Langston, A., Sebastion, D., and Wilson, G. (2005). The Limnology o f Williston Reservoir: British Columbia’s Largest Lacustrine Ecosystem. Water Quality Research Journal o f Canada, 40(1), 28-50. Storelli, M.M., Barone, G., Piscitelli, G., and Marcotrigiano, G.O. (2007). Mercury in fish: concentration vs. fish size and estimates o f mercury intake. Food Additives and Contaminants, 12, 1353-1357. Tan, M.L., Ujihara, A., Kent, L., Hendrickson, I. (2011). Communicating Fish Consumption Advisories in California: What Works, What Doesn’t. Risk Analysis, 31(7), 1095-1106. The Province (2003). Northern natives to sue over mercury poisoning. Written by Suzanne Fournier, published May 28, 2003. TourismBC (2008). Northern British Columbia. Retrieved from http://www.hellobc.com on March 26, 2008. Turyk, M.E., Bhavsar, S.P., Bowerman, W., Boysen, E., Clark, M., Diamond, M., Mergler, D., Pantazopoulos, P., Schantz, S., and Carpenter, D.O. (2012). Risks and Benefits o f Consumption o f Great Lakes Fish. Environmental Health Perspectives, 120(1), 11-18. UNEP (United Nations Environment Program) (2008). The Global Atmospheric Mercury Assessment: Sources, Emissions, and Transport. Retrieved from http://www.chem.unep.ch/mercurv/Atmospheric Emissions/UNEP%20SUMMA RY%20REPQRT%20%20CORRECTED%20Mav09%20%20final%20for%20WEB%202008.pdf on September 11, 2012. U.S. EPA (U.S. Environmental Protection Agency) (2007). TEACH Chemical Summary. Retrieved from http://www.epa.gov/teach/chem summ/mercury org summarv.pdf on September 11 , 2 0 1 2 . U.S. EPA (U.S. Environmental Protection Agency) (2011). Integrated Risk Information System, Methylmercury. Retrieved from http://www.epa.gov/iris/subst/0073.htm on September 11, 2012. U.S. Fish and Wildlife Service (1998). Bull Trout Facts. Retrieved from http://librarv.fws.gov/Pubsl/bulltrout.pdf on April 26, 2012. Wang, S., Biao, L., Zhang, M., Xing, D., Jia, Y., and Wei, C. (2012). Bioaccumulation and trophic transfer of mercury in a food web from a large, shallow, hypereutrophic lake (Lake Taihu) in China. Environmental Science and Pollution Research, 19(7), 2820-2831. 77 Weech, S. A., Scheuhammer, A. M., Elliott, J. E., & Cheng, K. M. (2004). M ercury in fish from the Pinchi Lake Region, British Columbia, Canada. Environmental Pollution, 131(2), 275-286. WHO (World Health Organisation) (2006). Elemental Speciation in Human Health Risk. Obtained from: http://www.who.int/ipcs/publications/ehc/ehc234.pdf on September 11, 2012. Wilson, R. (2010). Human Health Risk Assessment o f Pinchi Mine and Pinchi Lake Area. Final Report. Prepared for Teck Metals Ltd. by Wilson Scientific Consulting. 78 10.0 Appendices Table 10.1 Descriptive statistics fo r fish param eters used in this study Std. Skewness Skewness Kurtosis Kurtosis Std. Std. Error Deviation Error 0.044 195.955 0.088 3097 0.01 39.4 1.2829 1.62532 9.309 0.058 4.585 0.115 1801 27 1800 435.83 165.242 0.928 N Min Max Mean Weight (kg) Length (mm) Hg Concentration (ppm) 3097 0 8.31 0.3013 Year 3097 1974 2000 1983.83 Location 3097 1 34 19.54 0.54556 7.413 11.829 79 8.072 1.03 -0.356 0.044 93.424 0.044 -0.014 0.044 -1.406 0.088 0.088 0.088 Table 10.2 Descriptive statistics fo r total number offish used in this study by w ater body Frequency Percent Valid1 Percent 9.7 9.7 Babine Lake 301 Bear Lake 102 3.3 3.3 0.2 Brown Lake 5 0.2 Chuchi Lake 3.1 3.1 95 Cunningham Lake 1.8 1.8 57 0.6 0.6 Francois Lake 19 0.9 0.9 28 Grassham Lake 2.1 2.1 Inzana Lake 65 Kazchek Lake 1.5 1.5 46 1.1 1.1 Kemess Lake 35 0.3 0.3 McKnight Lake 8 0.3 0.3 8 Nations Lakes 0.5 0.5 Nechako Reservoir - Tahtsa Reach 16 Necoslie River - Stuart Lake 0.8 0.8 26 5.2 Pinchi Lake 162 5.2 0.6 20 0.6 Purvis Lake 3.6 110 3.6 Quesnel Lake 0.4 0.4 Rainbow Creek 11 4.6 Stuart Lake 141 4.6 34 1.1 1.1 Takla Lake 2.3 Tatchi River 72 2.3 2 62 2 Tchentlo 6.2 6.2 Tezzeron Lake 191 0.4 0.4 Tochcha Lake 11 3.4 3.4 Trembleur Lake/Middle River 105 6.2 6.2 191 Tsayata Lake 1.3 40 1.3 Weisner Lake 2.1 2.1 Whitefish Lake 66 334 10.8 Williston Reservoir - Finlay 10.8 5.8 Williston Reservoir - Ingenika 180 5.8 1.4 Williston Reservoir - Williston Lake 42 1.4 5.4 166 5.4 Williston Reservoir - Parsnip 9.8 304 9.8 Williston Reservoir - Peace 1.4 44 1.4 Witch Lake 100 100 Total 3097 1 V alid percen t w as used in the SPS S p ro g ra m to d istin g u ish w h e th e r th e re w e re any p a ra m e te rs m issin g . It is irrele v an t in th is tab le; how ever, in th e ta b le s th a t re fe r to le n g th o f fish it re fe rs to th o se fish th a t h ad m issing len g th s in the o riginal data so u rce. 80 Table 10.3 Descriptive/tests o f normality f o r fish param eters used in this study Weight (kg) Hg Concentration (ppm) Mean 95% Confidence Interval for Mean Lower Bound Upper Bound 1.3402 1.1093 0.75 2.642 1.62532 0.01 39.4 39.39 1.48 0.044 9.309 195.955 0.088 0.3013 0.00976 5% Trimmed Mean Median Variance Std. Deviation Minimum Maximum Range Interquartile Range Skewness Kurtosis Mean 95% Confidence Interval for Mean Std. Error 1.2829 0.02921 1.2256 Lower Bound Upper Bound 5% Trimmed Mean Median Variance Std. Deviation Minimum Maximum Range Interquartile Range Skewness Kurtosis 0.2822 0.3205 0.2225 0.17 0.295 0.54336 0 8.31 8.31 0.24 8.072 93.424 0.044 0.088 Table 10.4 Tests o f normality fo r fish weight and Hg concentrations Weight (kg) Hg Concentration (ppm) Shapiro-Wilk Ko lmogoro v- S mimo va Statistic d f Sig. df Sig. Statistic 0 0.217 3097 0.586 3097 0 0.407 3097 0.29, 3097, 0.00 0 81 Table 10.5 Square root transformation o f weight o f a ll fish used in this study sqrtWeight Mean 95% Confidence Interval for Mean Lower Bound Upper Bound 5% Trimmed Mean Median Variance Std. Deviation Minimum Maximum Range Interquartile Range Skewness Kurtosis Std. Error 0.00971 0.9954 0.9763 1.0144 0.9636 0.866 0.292 0.54058 0.08 6.28 6.19 0.77 0.044 1.315 0.088 5.709 Table 10.6 Tests o f normality fo r weight o f all fish used in this study sqrtWeight Kolmogorov-Smimova Statistic d f Sig. 0.122 3097 0 Shapiro-Wilk Statistic 0.92 d f Sig. 0 3097 Table 10.7 Hg concentration (ppm) with cube root transformation for fish in this study Mean 95% Confidence Interval for Mean Lower Bound Upper Bound 5% Trimmed Mean Median Variance Std. Deviation Minimum Maximum Range Interquartile Range Skewness Kurtosis 82 Std. Error 0.5836 0.00398 0.5758 0.5914 0.5671 0.554 0.049 0.22142 0 2.03 2.03 0.25 0.044 1.549 5.378 0.088 Table 10.8 Tests o f normality fo r H g concentration f o r all fish used in this study Kolmogorov-Smimova Statistic 0.094 df Sig. 0 3097 Shapiro-Wilk Statistic 0.905 df Sig. 0 3097 Table 10.9 Interlake differences ofH g concentrations (ppm) in fish tissue, sorted by species Fish Type Kokanee Lake trout Sockeye salmon Kokanee Lake trout Whitefish Lake whitefish Lake trout Lake whitefish Rainbow trout Burbot Lake trout Lake whitefish Lake whitefish Lake trout Bull trout Lake whitefish Bull trout Lake whitefish Bull trout Lake whitefish Rainbow trout Bull trout Lake whitefish Location Babine Lake Bear Lake Chuchi Lake Cunningham Lake Kazchek Lake Pinchi Lake Quesnel Lake Tatchi River Tezzeron Lake Tsayata Lake Whitefish Lake Williston Reservoir - Finlay Williston Reservoir - Ingenika Williston Reservoir - Parsnip Williston Reservoir - Peace 83 M ean 0.0555 0.2422 0.0514 0.0315 0.3268 0.2718 0.0502 1.8218 0.4954 0.1165 0.2795 0.6064 0.0895 0.1876 0.3108 0.846 0.1873 0.7027 0.2251 0.5135 0.1895 0.0415 0.4679 0.1299 N 71 108 93 40 59 49 43 75 51 73 44 84 65 153 40 160 107 45 57 43 47 46 57 158 Table 10.10 Mean Hg concentrations (ppm) in fish tissue in various fish species Species Bull trout Burbot Char Dolly varden Kokanee Lake trout Lake whitefish Mountain whitefish Rainbow trout Sockeye salmon Whitefish Total M ean 0.7021 0.3092 0.377 0.5215 0.0845 0.5017 0.1681 0.0989 0.1209 0.0535 0.2368 0.2886 N 313 137 66 87 195 660 892 60 365 147 81 3003 Table 10.11 Hg concentration in fish tissue collected in various years Year 1974 1975 1976 1977 1978 1979 1980 1981 1985 1986 1988 1989 1992 1993 1996 2000 Total Std. Deviation Mean N 42 2.1148 0.48 6 0.2456 160 0.1915 234 0.1746 275 0.2374 675 0.3763 303 0.1826 151 0.412 5 0.8414 58 0.3032 750 0.0856 11 0.0277 35 0.0869 16 0.3662 13 0.334 363 0.3013 3097 % of total N 2.7941 1.4% 0.17967 .2% 0.15967 5.2% 0.17481 7.6% 0.13862 8.9% 0.22943 21.8% 0.35333 9.8% 0.25814 4.9% 0.04764 .2% 0.56535 1.9% 0.54715 24.2% 0.03875 .4% 0.05096 1.1% 0.18297 .5% 0.24199 .4% 0.45972 11.7% 0.54336 100.0% 84 Variance 7.807 0.032 0.025 0.031 0.019 0.053 0.125 0.067 0.002 0.32 0.299 0.002 0.003 0.033 0.059 0.211 0.295 Table 10.12 Hg concentration (ppm) in all fish sam pled from each location Lake Tsayata Lake Babine Lake Bear Lake Brown Lake Chuchi Lake Cunningham Lake Weisner Lake Whitefish Lake Francois Lake Grassham Lake Witch Lake Inzana Lake Kazchek Lake Kemess Lake McKnight Lake Trembleur Lake/Middle River Nations Lakes Nechako Reservoir - Tahtsa Reach Necoslie River - Stuart Lake Stuart Lake Pinchi Lake Purvis Lake Quesnel Lake Rainbow Creek Tatchi River Takla Lake Tchentlo Tezzeron Lake Tochcha Lake Williston Reservoir - Finlay Williston Reservoir - Ingenika Williston Reservoir - Williston Lake Williston Reservoir - Parsnip Williston Reservoir - Peace Total Mean N Std. Deviation % of total N Variance 0.15074 6.2% 0.023 0.228 191 0.013 0.1316 301 0.1159 9.7% 0.024 0.1228 102 0.156 3.3% 0.112 0 5 0.02049 .2% 0.04 0.3052 95 0.20086 3.1% 0.021 0.2602 57 0.14599 1.8% 0.016 0.1915 40 0.12771 1.3% 0.029 0.2473 66 0.1715 2.1% 0.016 0.1633 19 0.12811 .6% 0.004 0.06551 .9% 0.1739 28 0.14002 1.4% 0.02 0.2575 44 0.08541 2.1% 0.007 0.1272 65 0.04164 1.5% 0.002 0.0533 46 0.0277 35 0.05096 1.1% 0.003 0.025 0.525 0.15866 .3% 8 0.1452 105 0.13014 3.4% 0.017 0.02 0.2725 0.14109 .3% 8 0.033 0.18297 .5% 0.0869 16 0.3042 26 0.2449 .8% 0.06 0.02 0.14005 4.6% 0.1606 141 2.664 1.1671 162 1.63217 5.2% 0.012 0.1775 20 0.11045 .6% 0.13462 3.6% 0.018 0.1577 110 0.002 0.0856 11 0.03875 .4% 0.021 72 0.14614 2.3% 0.2781 0.014 0.1762 34 0.11626 1.1% 0.02 0.2065 62 0.1401 2.0% 0.3517 191 0.31616 6.2% 0.1 0.13372 .4% 0.018 0.51 11 0.5244 334 0.72657 10.8% 0.528 0.4335 180 0.38227 5.8% 0.146 0.13056 1.4% 0.017 0.2579 42 0.112 0.2558 166 0.335 5.4% 0.3281 9.8% 0.1829 304 0.108 0.54336 100.0% 0.3013 3097 0.295 85 Table 10.13 Mean weight (kg) o f fish, sorted by species Fish Type Bull trout Burbot Char Coho salmon Dolly varden Kokanee Lake trout Lake whitefish Large scale sucker Mountain whitefish Peamouth chub Rainbow trout Rocky mt. whitefish Salmon Sockeye salmon Squawfish Sturgeon Sucker Trout White sucker Whitefish Total Mean 1.5366 1.1503 2.4091 1.5 1.005 1.3796 2.2005 0.4547 1.175 0.8973 0.2037 1.061 0.5 2.1846 2.035 1.5 39.4 0.7196 0.3339 0.7045 0.8185 1.2829 86 N Median 313 137 66 4 87 195 660 892 3 60 3 365 1 13 147 1 2 23 33 11 81 3097 1 1.2 2.1 1.5 0.5 1.5 1.9 0.4 1.15 0.8 0.21 0.4 0.5 2 2 1.5 39.4 0.5 0.3 0.5 0.8 0.75 Table 10.14 Mean weight (kg) o f fish, sorted by location Location Tsayata Lake Babine Lake Bear Lake Brown Lake Chuchi Lake Cunningham Lake Weisner Lake Whitefish Lake Francois Lake Grassham Lake Witch Lake Inzana Lake Kazchek Lake Kemess Lake McKnight Lake Trembleur Lake/Middle River Nations Lakes Nechako Reservoir - Tahtsa Reach Necoslie River - Stuart Lake Stuart Lake Pinchi Lake Purvis Lake Quesnel Lake Rainbow Creek Tatchi River Takla Lake Tchentlo Tezzeron Lake Tochcha Lake Williston Reservoir - Finlay Williston Reservoir - Ingenika Williston Reservoir - Williston Lake Williston Reservoir - Parsnip Williston Reservoir - Peace Total Mean 0.6076 1.8967 1.6426 0.1062 2.2763 0.8754 1.7125 2.1606 1.2618 0.2036 1.1159 1.3662 0.7663 0.1785 0.3696 1.0395 2.0687 0.2039 2.4923 1.9585 1.5108 0.505 2.8727 0.0512 1.7014 1.0574 1.45 1.481 3.3636 1.0578 0.9292 0.3688 0.6482 0.5778 1.2829 87 N 191 301 102 5 95 57 40 66 19 28 44 65 46 35 8 105 8 16 26 141 162 20 110 11 72 34 62 191 11 334 180 42 166 304 3097 Median 0.5 1.9 1.5 0.078 2 0.8 1.6 1 0.775 0.2 1 1.4 0.8 0.032 0.372 0.5 1.625 0.194 2.15 1.25 0.75 0.35 2.75 0.039 1.5 1 0.8875 1.15 3.5 0.4675 0.5 0.2 0.35 0.35 0.75 Table 10.15 Hg concentration offish tissue, sorted by year, lake, and fish species. Year Location Fish Type 1974 Pinchi Lake Kokanee Lake trout Lake whitefish Large scale sucker Mountain whitefish Peamouth chub Rainbow trout Total 1976 Necoslie River - Stuart Lake Char Salmon Sucker Total Stuart Lake Sturgeon Total 1979 Tezzeron Lake Burbot Lake trout Lake whitefish Rainbow trout 88 Mean N Mean N Mean N Mean Weight Hg Concentration (ppm) (kg) 0.48 0.1625 6 6 4.8275 2.4188 12 12 0.3 1.79 5 5 2.5867 1.175 N Mean 3 0.66 3 0.365 N Mean N Mean N Mean N Mean 2 1.8967 3 0.39 11 2.1148 42 0.4971 2 0.2037 3 0.3023 11 0.945 42 2.8357 N Mean N Mean N Mean N Mean N Mean N Mean N Mean N Mean N Mean N 14 0.13 1 0.3 1 0.4619 16 0.45 2 0.45 2 0.455 12 0.6309 66 0.1012 32 0.1676 29 14 4.1 1 1.3 1 2.8188 16 39.4 2 39.4 2 1.8333 12 2.2333 66 0.5606 32 0.9 29 Whitefish Mean N Mean N Mean 0.11 1 0.3951 140 0.6379 0.5 1 1.5281 140 1.2179 N Mean N Mean N Mean N Mean N Mean 14 0.26 20 0.189 30 0.26 10 0.3027 74 0.5815 14 0.66 20 0.3583 30 0.66 10 0.6432 74 0.6885 N Mean N Mean N Mean N Mean N Mean N Mean 26 0.5664 14 0.1813 30 0.07 1 0.29 1 0.4007 72 0.7027 26 0.675 14 0.4987 30 0.25 1 0.6 1 0.5994 72 1.5693 Bull trout N Mean N Mean N Mean N Mean N Mean 45 0.7324 21 0.2062 37 0.3333 3 0.5248 106 0.784 45 1.5381 21 0.4197 37 0.4 3 1.1288 106 4.22 Dolly varden N Mean 5 0.6807 5 3.2167 Total Williston Reservoir Ingenika Dolly varden Lake whitefish Rainbow trout Whitefish Total 1980 Williston Reservoir Finlay Bull trout Dolly varden Lake whitefish Rainbow trout White sucker Total Williston Reservoir Ingenika Bull trout Dolly varden Lake whitefish White sucker Total 1981 Williston Reservoir Finlay 89 1986 Pinchi Lake Tochcha Lake 1988 Williston Reservoir Finlay____________ Williston Reservoir Parsnip___________ Lake whitefish Mean 0.146 0.46 White sucker Mean 0.31 0.4167 Total Mean 0.4331 1.9271 Bull trout Mean 0.6533 0.36 Lake trout Mean 1.1033 2.2373 Lake whitefish Mean 0.645 0.84 Rainbow trout Mean 0.195 0.3475 Total Mean 0.9189 1.6803 Lake trout Mean 0.51 3.3636 Total Mean 0.51 3.3636 Bull trout Mean 1.1369 2.1345 Burbot Mean 0.3325 0.3 Kokanee Mean 0.1922 0.254 Lake whitefish Mean 0.2795 Rainbow trout Mean 0.2277 44 0.077 Total Mean Bull trout Mean 0.6585 170 0.5135 0.4129 24 1.2144 170 1.2752 Burbot Mean 0.3194 0.8493 Lake trout Mean 0.315 0.595 Lake whitefish Mean 0.1895 0.2621 90 Rainbow trout Total 1996 McKnight Lake Dolly varden Total 2000 Pinchi Lake Lake trout Lake whitefish Total Tezzeron Lake Lake trout Lake whitefish Total 91 Mean N Mean N Mean N Mean N Mean N Mean N Mean N Mean N Mean N Mean N 0.0415 46 0.2558 166 0.525 8 0.525 8 1.3821 33 0.2522 34 0.8087 67 0.5165 18 0.0781 33 0.2328 51 0.3365 46 0.6482 166 0.3696 8 0.3696 8 3.3153 33 0.4618 34 1.8673 67 2.8917 18 0.5114 33 1.3515 51 Table 10.16 Latin names fo r fish species used in this study Common Name Bull trout Burbot Char Coho salmon Dolly varden Kokanee Lake trout Lake whitefish Large scale sucker Mountain whitefish Peamouth chub Rainbow trout Rocky mt. whitefish Salmon Sockeye salmon Squawfish Sturgeon Sucker Trout White sucker Whitefish L atin Name Salvelinus confluentus Lota lota Salvelinus fontinalis Oncorhynchus kisutch Salvelinus malma malma Oncorhynchus nerka Salvelinus namaycush Coregonus clupeaformis Catostomus macrocheilus Prosopium williamsoni Mylocheilus caurinus Oncorhynchus mykiss Prosopium williamsoni Salmo Salar Oncorhynchus nerka Ptychochelius Acipenser sturio Catostomus commersonii Salvelinus malma malma Catostomus commersonii Coregonus clupeaformis 92 Figure 10.1 Scatterplot o f Hg concentration (ppm) versus weight (kg) fo r bull trout y = 0.371 x + 0.132 5.00- 3.00- X + .1 33 R Sq Linear = 0.571 1 .00 - 0 ,000.00 2.00 4.0Q 6.0Q 8.00 10.00 Weight (kg) Figure 10.2 Sample calculation fo r adjusted Hg concentration (ppm) of fish fo r the Risk Assessment Tool Calculation fo r a bull trout with an actual concentration o f 0.07 ppm and a weight o f 0.18 kg A d ju sted P P M = (actual p p m ) x (m ean sp e c ie s p p m /c a lc u la te d p p m acco rd in g to re g re ssio n ) = (0 .0 7 ) x (0 .7 0 2 1 /((0 .3 7 1 x 0 .1 8 ) + 0 .1 3 2 ) = (0 .0 7 ) x (0 .7 0 2 1 /0 .1 9 8 7 8 ) = (0 .0 7 ) x (3 .5 3 2 0 4 ) = 0 .2 4 7 2 4 93