Assessing the sublethal effects of in-river concentrations of parameters contributing to cumulative effects in the athabasca river basin using a fathead minnow bioassay

Authors


Abstract

The Athabasca River basin, located in Alberta, Canada, covers 157, 000 km2 and holds significant cultural and economic importance. Recent research assessed changes in several water quality and quantity parameters that have changed both spatially (along the river continuum) and temporally (pre-development and present day) in the Athabasca River Basin. In particular, parameters such as salinity and dissolved sulphate have changed significantly across the Athabasca River mainstem over the past five decades. Further laboratory testing has linked concentrations of these parameters to changes in fathead minnow reproduction. Research is required to determine whether these changes observed in the laboratory can be applied to actual in-river conditions. The objectives of the present study were to twofold: assess changes in fathead minnow response metrics (i.e., condition, liver and gonad size, egg production, and gill histology) associated with increasing concentrations of salinity and dissolved sulphate and determine whether sublethal effect thresholds established in laboratory experiments correspond to actual in-river concentrations using water from the mouth and headwaters of the Athabasca River. Three dose−response experiments (NaCl, SO4, and water sampled from the mouth of the Athabasca River) were conducted at Jasper National Park, Alberta, Canada. Significant increases in mean eggs per female per day occurred at the 50% treatment for the mouth experiment and thresholds previously developed in the laboratory were verified. Environ. Toxicol. Chem. 2013;32:662–672. © 2012 SETAC

INTRODUCTION

The Athabasca River Basin is located in Alberta, Canada and accounts for approximately 22% of the landmass of Alberta 1. It originates at the Columbia Ice Fields in Jasper National Park and flows northeast 1,300 km across Alberta until it terminates in Lake Athabasca. The Athabasca River has experienced an increasing level of development related to land use over the past five decades, including forestry/pulp and paper, coal mining, oil and natural gas, agriculture, tourism, wildlife trapping, hunting, and oil sands mining 2, 3. As a result, several studies have examined portions of the basin, which have identified some water quality and quantity issues of particular concern 2, 4.

Various governmental and non governmental organizations have conducted independent monitoring programs in this river basin over the past few decades. Such efforts provide the opportunity to integrate information in a cumulative effects context at a watershed scale. It is important to assess cumulative effects over large space and time scales, because they can occur for multiple reasons including natural phenomena, industrial growth, population growth and economic development.

Several oil sands mining operations are located near the mouth of the Athabasca River in the Wood Buffalo region of Alberta, Canada. As part of the extraction process, vast amounts of slurry called oil sands process water (OSPW) is produced and stored onsite in large lakes and holding ponds 5. This OSPW contains several inorganic and organic constituents, some of which have been shown to be acutely toxic to aquatic life. Some of these constituents include high levels of salinity, sulphate, ammonia, conductivity, and naphthenic acids 6. There is the potential for some of these components to make their way into the Athabasca River through seepage and air deposition 7. Previous research has shown that the highest concentrations of sodium and chloride along the Athabasca River mainstem occurred in the mouth reach of the river basin 8. Assessing water directly from the mouth of the Athabasca River, downstream of all major inputs, would help determine the current and potential future impact these developments can have on aquatic biota.

The oil sands deposits are thought to be of marine origin 9, 10. The dominant ions in seawater are chloride and sodium, with other major ions such as sulfate, calcium, and potassium 11. As a result, salinity is becoming an increasing concern to the oil sands industry and its regulators in northern Alberta 12. Recently, it has been noted that dissolved sodium, chloride, and sulphate concentrations have increased significantly in the lower reaches of the Athabasca River over 40 years 8.

Increases in salinity have been shown to cause both acute and chronic effects at specific life stages in many fish species. This can result in excluding less-tolerant species, which can limit biodiversity and cause a shift in biotic communities 13. To acclimatize successfully to a new salinity involves a great deal of energy to engage several physiological (in the gills, intestine, and kidney) and behavioral responses 14. This can often come at the expense of other processes, such as growth and reproduction. Laboratory exposures with both sodium and chloride have been shown to alter reproduction in fathead minnow and gill membrane structure 15.

In most freshwater systems, sulphate is found in very low concentrations, with the exception being in areas where sulphate-containing ores or anthropogenic activities occur 16. Anthropogenic activities, which can contribute sulphate include mining, smelting, burning fossil fuels, agricultural runoff, and domestic sewage and may account for up to 90% of the sulphur found in surrounding aquatic systems 16. Sulfate, therefore, is also an ion of interest in the lower Athabasca region.

Based on present development along the Athabasca River and the results of previous studies, there is a need to investigate the effects that exposure to sodium, chloride, and sulphate at chronic (lower) levels have on the aquatic biota. At this time, it is unknown whether the levels of these ions currently present in the mouth of the Athabasca River are affecting the health and reproduction of the aquatic biota living in the area.

The test species chosen for use in the present study is the fathead minnow (Pimephales promelas). It is a freshwater species that is widely distributed across North America. It is easy to raise and breed under laboratory conditions due to its relatively rapid life cycle. Consequently, it is a species that is commonly used in standard toxicity testing. Furthermore, several protocols have been developed for culturing, handling, and toxicity testing 17, 18. The objectives of the present study were to twofold. First, it sought to verify changes in fathead minnow (P. promelas) indicators as seen in previous laboratory studies with increasing concentrations of sodium chloride and dissolved sulphate that are deemed to be important to the Athabasca River. Second, the present study was undertaken to determine the sublethal effects on P. promelas indicators using water from both the headwaters and mouth of the Athabasca.

METHODS

To assess potential changes in fish reproduction P. promelas, partial lifecycle bioassay was used. This assay allowed us to assess the reproduction of fathead minnows and aspects of their early development in a timeframe much shorter than a traditional lifecycle bioassay. It is based on partial lifecycle tests originally developed by Ankley et al. 17 and further refined by Rickwood 19. All experiments were conducted at the headwaters of the Athabasca River mainstem in Jasper National Park, Jasper, Alberta, Canada using the The Healthy River Ecosystem Assessment System (THREATS) trailer from June to August 2009. The trailer was held under controlled conditions (16:8 light:dark photoperiod and water temperature equal to 25°C ± 1°C).

Water chemistry

To expose fish to test solutions, a diluter system was used that allowed for a six-time dilution of a 100% test solution with three replicates per dilution. The 100% concentrations chosen for the sodium chloride and dissolved sulphate experiments were based on the average levels of these parameters found along the river continuum of the Athabasca River as published in previous research 8; therefore, they were ecologically relevant to the Athabasca River Basin. The dilutions series used were 0% (control), 6.25, 12.5, 25, 50, and 100% (25 ppm Na/38 ppm Cl). The salt NaCl was used to generate the 100% test solution for the sodium chloride experiment, and the salt NaSO4 was used to generate the 100% test solution for the dissolved sulphate experiment. To obtain enough water to conduct the mouth water experiment, 1,800 L of water was collected from the Athabasca River mainstem downstream of major oil sand developments and trucked to the experimental site in Jasper National Park. Control water was taken directly from the headwaters of the Athabasca River at the experimental site. Each test tank (replicate) had a complete turnover of water four times daily.

Samples of water from each treatment were collected daily during the exposure period and analyzed for general chemistry parameters. These included dissolved oxygen, temperature and conductivity (YSI portable meter, Yellow Springs Instrument), pH (Oakton phTestr30), hardness (Hatch Test Kit Model 5-EP MG-L), and ammonia (Hagen Ammonia Test Kit A7820). In addition, each treatment was sampled on a weekly basis and submitted to ALS Laboratories and analyzed for dissolved sodium (American Public Health Association [APHA] Method 3120 B-ICP-OES); dissolved chloride (APHA Method 4500 CL-E); total kjeldahl nitrogen (APHA Method 4500N-C-Dig.-Auto-Colorimetry); total organic carbon (APHA Method 5310 B-Instrumental); ammonia (APHA Method 4500-NH3-G); total phosphorous (APHA Method 4500-P-B, C Auto-Colorimetry); total suspended solids (APHA Method 2540 D-Gravimetric); sulphate (APHA Method 3120 B-ICP-OES); and turbidity (APHA Method 2130 B).

Fish reproduction

The present experiments were conducted using 6- to 9-month-old fathead minnows obtained from Osage Catfisheries. The present study was approved by the University of Saskatchewan's Animal Research Ethics Board and adhered to the Canadian Council on Animal Care guidelines for humane animal use. The first phase of these experiments was the pre-exposure phase, which lasted 7 to 10 d. During this phase, approximately twice the number of breeding trios (1:2 male:female) of fathead minnows required for the exposure phase were each placed into a 9 L aquarium with a breeding tile and an air stone. Fish were fed frozen brine shrimp and bloodworms twice daily. The secondary sexual characteristics, total body weight, and total length of each fish were recorded prior to being added to the test tanks. Previous research has suggested that length ratio of the male fish to females is a good indicator of successful breeding; therefore, each of the females was length matched (ideally 75% of the length of the male) 20. Every 24 h, each breeding tile was checked for the presence of eggs. If eggs were found, they were scraped off the tile and photographed for counting. After this phase, three breeding trios that meet test criteria (minimum of 80% fertilization, bred at least once, and adults survived the entire pre-exposure period) were assigned randomly to each treatment. Mean egg production and fertilization success of each treatment before exposure were tested using a one-way analysis of variance (ANOVA) providing assumptions of normality (Shapiro-Wilks) and homogeneity of variance (Levene's) were met. For data that did not meet these assumptions, the non parametric equivalent (Kruskal-Wallis test) was used. This was done to ensure treatments were not statistically different from each other (p > 0.05). Therefore, prior to exposure, each treatment had the potential for similar egg production.

The second phase of these experiments was an exposure phase lasting 21 d. During the exposure period, each of the breeding trios that were assigned randomly in the pre-exposure phase was exposed to a particular concentration of dissolved sodium or dissolved chloride (see Table 1). Throughout the exposure phase, breeding tiles were checked daily for eggs, which were removed from the tile and photographed for counting. They were then placed in PVC cups with a mesh bottom and air stone and placed in separate tanks filled with the same concentration of exposure water as their parents. Eggs were checked after 48 h and then daily thereafter, and the stages of development were recorded (eyed, first hatched, fully hatched). Eggs that were eyed were considered fertilized, and the batch of eggs was then photographed for counting (percentage fertilized). The total number of eggs, number of fertilized eggs, time to hatch, and number of larvae (alive, deformed, and dead) after 5 d posthatch were recorded. After 5 d posthatch, larvae were anaesthetized using methane tricainesulfonate (MS222, ∼1,000 mg/L) and preserved in 10% formalin. After 21 d of exposure, the adult fish were anaesthetized using methane tricainesulfonate (MS222, ∼1,000 mg/L), and then euthanized using spinal severance prior to further processing. Secondary sexual characteristics, total body weight, total length, carcass weight, liver weight, and gonad weight were recorded. The second gill arch was removed from the right side of the fish and preserved in 10% formalin for further histological examination.

Table 1. Dissolved sodium, dissolved chloride, dissolved sulphate, total organic carbon (TOC), and turbidity averaged for each treatment across the duration of the exposure period (21 d) of each experiment. Mean ± (standard deviation)
ExperimentDilution treatmentDissolved sodium (mg/L)Dissolved chloride (mg/L)Dissolved sulphate (mg/L)Total organic carbona (mg/L)Turbiditya (NTU)
  • a

    Measurements for TOC and turbidity in the 6.25, 12.5, 25, 50, and 100% treatments are single measurements made during the first week of the experiment.

  • b

    Statistical significance = p ≤ 0.05 versus 0% control.

  • ND = not detected.

Sodium chloride0%1.15 (0.15)ND12.73 (1.29)1.30 (0.30)4.80 (1.40)
6.25%3.67 (0.58)4.30 (0.59)12.53 (1.34)1.305.10
12.5%3.23 (0.30)4.07 (0.20)12.53 (1.41)14.805.15
25%8.07 (0.26)10.83 (0.13)12.50 (1.21)ND5.94
50%11.37 (1.09)16.63 (1.27)12.80 (1.15)1.206.04
100%25.43 (0.68)38.90 (0.15)13.10 (1.08)2.104.42
Dissolved sulphate0%1.70 (0.50)ND15.03 (0.93)2.80 (1.60)3.29 (0.86)
6.25%3.67 (0.20)ND18.57 (1.37)1.305.13
12.5%6.13 (0.15)ND23.33 (1.79)1.105.07
25%11.00 (0.31)ND34.43 (1.91)1.804.07
50%22.00 (0.72)ND57.13 (3.10)1.203.87
100%43.23 (1.22)ND101.37 (4.88)6.903.28
Mouth water0%2.10 (0.40)1.40 (0.40)13.13 (1.89)2.00 (0.40)3.19 (0.38)
6.25%1.87 (0.28)ND12.97 (1.63)3.07 (0.61)3.57 (0.93)
12.5%2.50 (0.26)1.53 (0.18)15.30 (2.81)4.27 (1.17)b3.27 (0.13)
25%3.80 (0.17)2.30 (0.17)13.67 (1.12)4.80 (0.15)b3.50 (0.39)
50%6.20 (0.31)4.07 (0.09)13.83 (0.71)8.20 (0.38)b5.58 (1.22)
100%8.23 (2.52)5.73 (2.02)14.40 (0.61)10.73 (3.62)b4.11 (0.68)

Gill histology

Gill arches were submitted for further processing to Prairie Diagnostic Services (Saskatoon, Canada). Each arch was embedded using routine paraffin processing techniques and then sliced. Two to three slices were placed on a slide and then stained using a standard hematoxylin eosin stain. Following staining, they were covered for further examination. Slides were examined using an Axio Observer Z1 microscope (Carl Zeiss MicroImaging GmbH) and photographed using an AxioCam ICc1. Prior to examination, all slides were randomly assigned a number to remove treatment bias. One photomicrograph was taken of each gill arch and the third primary lamella from the left side of each arch was selected for further measurement.

On the right side of this primary lamella, three secondary lamellae at the proximal, three in the middle, and three at the distal areas were measured for secondary lamellar length (SLL) and secondary lamella width (SLW). These three areas were determined in each individual fish based on the total length of the primary lamella (i.e., the three middle secondary lamella were located at 50% of the total length of the primary lamella). In addition, one measurement of the basal epithelium thickness (BET) was taken at each of these three points along the secondary lamella. To quantify the diffusion distance across the gills further, an additional endpoint was calculated using the average SLW and average SLL for each gill arch to calculate the ratio of width and length of each arch (SLW/SLL). A high ratio of SLW to SLL would mean that the lamella is wide and short, thereby creating a larger distance for ion diffusion across the gill. A low ratio of SLW to SLL, however, would mean that the lamella is thin and long, which would reduce the distance required for ion diffusion across the gill.

The averages of BET on each gill arch and the endpoint SLW/SLL were analyzed using a one-way ANOVA providing assumptions of normality (Shapiro-Wilks) and homogeneity of variance (Levene's) were met. For data that did not meet these assumptions, the non parametric equivalent (Kruskal-Wallis test) was used. Differences were considered to be significant when p < 0.05.

Statistics

At the end of the exposure period, fish metrics, reproductive endpoints and larval endpoints were analyzed. All statistical analyses were performed using SPSS 17 and graphed using Sigmaplot Version 11. Results were considered significant when p < 0.05.

To assess cumulative frequency data, the Kolmogorov-Smirnov test was used. Endpoints included: cumulative eggs/female (cumulative number of eggs produced per treatment/number of living females/number of days), which factors in the effects of mortality on egg production and represents population effects over time and cumulative spawning events (cumulative total number spawning events/treatment/d]). One-way ANOVAs or non parametric equivalent (Kruskal-Wallis) tests were conducted on mean egg data. A one-way ANOVA or its nonparametric equivalent was conducted on the following endpoints: hatching success; percentage of deformities; liver somatic index (LSI) (liver wt [g]/body wt [g] × 100); gonadal somatic index (GSI) (gonad wt [g]/body wt [g] × 100); condition factor (body wt [g]/total length[cm]3 × 100); mean total egg production (total number of eggs produced per breeding group/number of females in group/number of exposure days); mean egg production (mean number eggs produced per stream/number of females in group/number of exposure days); and water quality. The one-way ANOVA was used when the data met assumptions (normal distribution and homogeneity of variance), which were analyzed using Levene's and Shapiro Wilk's tests. If data did not meet these assumptions, they were transformed (log transformation of continuous or derived data and arcsine transformation of percentage-based or ratio scaled data). If data still did not meet these assumptions, the non-parametric equivalent of the one-way ANOVA (Kruskal-Wallis test) was conducted. Differences among treatment groups were assessed further using a Dunnett's post hoc or non parametric Mann-Whitney U test.

RESULTS

Water chemistry

Dissolved oxygen, temperature, pH, ammonia, and conductivity measurements for all three experiments are listed in Table 2. All parameters were within acceptable limits for aquatic life throughout all three experiments 21. Significant increases in conductivity occurred in all three experiments. In the sodium chloride and dissolved sulphate experiments, 6.25, 12.5, 25, 50, and 100% treatments all had significantly higher conductivity than the 0% control. In the mouth-water experiment, significant increases in conductivity occurred in the 25, 50, and 100% treatments. Table 1 lists the levels of dissolved sodium, dissolved chloride, dissolved sulphate, total organic carbon, and turbidity in all three experiments. Levels of all three of these parameters were similar across all controls.

Table 2. General chemistry parameters (pH, temperature, dissolved oxygen, ammonia, and conductivity) averaged for each treatment across the duration of the exposure period (21 d) of each experiment. Mean ± (standard deviation)
ExperimentDilution treatmentDissolved oxygen (mg/L)Temperature (°C)pHAmmonia (mg/L)Conductivity (µS/cm)
  • a

    Statistical significance = p ≤ 0.001 versus 0% control.

Sodium chloride0%7.43 (0.88)20.16 (2.02)8.11 (0.13)0.15 (0.26)156.95 (5.72)
6.25%7.36 (0.69)20.06 (1.73)8.10 (0.12)0.15 (0.26)168.00 (11.13)a
12.5%7.15 (1.38)20.20 (1.76)8.09 (0.12)0.17 (0.44)171.61 (20.12)a
25%7.74 (0.56)20.16 (1.52)8.10 (0.09)0.16 (0.30)196.32 (23.05)a
50%7.52 (1.10)20.08 (1.84)8.14 (0.10)0.07 (0.12)212.29 (9.78)a
100%7.42 (1.16)19.43 (1.91)8.12 (0.09)0.27 (0.51)292.04 (14.77)a
Dissolved sulphate0%7.04 (1.07)22.15 (1.58)8.23 (0.20)0.13 (0.25)166.23 (22.39)
6.25%7.15 (0.53)22.69 (1.24)8.21 (0.15)0.15 (0.26)178.95 (19.15)a
12.5%6.73 (1.26)22.88 (1.32)8.16 (0.16)0.19 (0.28)185.78 (9.05)a
25%6.96 (1.06)22.87 (1.33)8.15 (0.13)0.10 (0.17)206.95 (10.74)a
50%6.83 (1.29)22.58 (1.55)8.13 (0.14)0.10 (0.16)261.44 (10.46)a
100%6.51 (1.32)22.48 (1.65)8.05 (0.16)0.07 (0.12)362.99 (23.46)a
Mouth water0%6.99 (1.32)19.92 (1.73)8.03 (0.17)0.27 (0.51)165.09 (17.24)
6.25%7.55 (0.87)20.12 (1.73)8.09 (0.13)0.12 (0.17)161.22 (5.52)
12.5%7.19 (1.09)20.18 (1.69)8.05 (0.13)0.11 (0.15)166.27 (4.55)
25%7.06 (1.32)20.46 (1.72)8.03 (0.11)0.12 (0.17)173.86 (4.15)a
50%6.78 (1.26)20.63 (1.56)8.03 (0.13)0.10 (0.12)184.97 (4.38)a
100%7.03 (0.78)21.20 (1.51)8.05 (0.15)0.16 (0.24)213.34 (10.28)a

The levels of dissolved sodium and chloride in the 100% treatment in the sodium chloride experiment were 25.43 and 38.90 mg/L, respectively. There was an unrepresentative spike in total organic carbon to 14.80 mg/L in the 12.5% treatment in the present experiment. There were no detectable levels of chloride in the dissolved sulphate experiment; however, there was 43.23 mg/L of dissolved sodium and 101.37 mg/L of dissolved sulphate measured in the 100% treatment of the present experiment. An increase in TOC was also observed in the 100% treatment of the dissolved sulphate experiment to 6.90 mg/L.

In the 100% treatment of the mouth-water experiment, the levels of dissolved sodium, dissolved chloride, and dissolved sulphate reached 8.23, 5.73, and 14.40 mg/L, respectively. There was a significant increase in total organic carbon in the 12.5, 25, 50, and 100% treatments compared to the 0% control. Although not significant, there was also an increase in turbidity in the 50% treatment to 5.58 nephelometric turbidity units (NTU).

Fish

Survival, condition factor, LSI, GSI, body weight, forklength, and secondary sexual characteristics in both male and female fish were not significantly different among treatments in the sodium chloride, dissolved sulphate, or mouth-water experiments. One female died in the 6.25% treatment on day 13 in the sodium chloride experiment and two females died in the 0% control in the dissolved sulphate experiment on days 18 and 20.

Reproduction

Hatching success, percentage of deformities, and cumulative spawning events showed no significant differences among treatments in the sodium chloride, dissolved sulphate, or mouth-water experiments (p > 0.05).

There were no significant differences among treatments for cumulative eggs per female, total eggs/female/day and mean eggs/female/d for the sodium chloride experiment compared to the 0% control (Figs. 1a and 2a). However, all treatments except the 12.5% treatment had higher cumulative egg production than the 0% control. Although not significant, only the 25% treatment had lower mean egg production than the 0% control in the sodium chloride experiment (Fig. 2a).

Figure 1.

Cumulative eggs per female for the 21-d exposure period in the (a) sodium chloride, (b) sulphate, and (c) mouth-water experiments. Concentrations of dissolved sodium and chloride used in each treatment are listed in Table 1.

Figure 2.

Mean eggs per female per day for the 21-d exposure period in the (a) sodium chloride, (b) sulphate, and (c) mouth-water experiments. Concentrations of dissolved sodium and chloride used in each treatment are listed in Table 1.

There was no difference among treatments for cumulative eggs/female, total eggs/female/day and mean eggs/female/day in the dissolved sulphate experiment compared to the 0% control (Figs. 1b and 2b). However at the end of the experiment (day 21 of exposure) all treatments had lower cumulative egg production than the 0% control (Fig. 1b). Although not significant, there was also an increase in mean egg production in the 25% treatment in the dissolved sulphate experiment (Fig. 2b).

There was no significant difference among treatments in the mouth experiment for cumulative eggs/female/day; however there was a significant increase in both total eggs/female/day and mean eggs/female/day (Figs. 1c and 2c) in the 50% treatment compared to the 0% control. Although not significant, all treatments (except the 25% treatment) had higher cumulative egg production than the 0% control (Fig. 1c).

The percentage of change from control for total eggs/female/day was calculated for all three experiments the (Fig. 3). In the sodium chloride experiment, the only negative changes were in the 6.25 (3.67 mg/L) and 12.5% (3.23 mg/L). In the dissolved sulphate experiment, three treatments were lower than the control, 6.25 (18.57 mg/L), 12.5 (23.33 mg/L), and 100% (101.37 mg/L). In the mouth-water experiment, the only negative change from control was in the 25% treatment (4.8 mg/L TOC). There was an increase of 175% in the 50% treatment (8.20 mg/L TOC).

Figure 3.

Percentage of change from control for total eggs per female per day for (a) sodium chloride, (b) dissolved sulphate, and (c) mouth water experiments.

Gills

There was no change in the BET in the sodium chloride experiment (Fig. 4b). However, there was a significant increase in SLW/SLL in the 12.5 (p < 0.001) and 100% (p < 0.05) treatments compared with the 0% control (Fig. 4a). In the dissolved sulphate experiment, there was a significant increase in BET in the 50% treatment (p = 0.039) (Fig. 5b) and significant decrease in SLW/SLL in the 6.25, 12.5, and 100% treatments (p < 0.05) compared with the 0% control (Fig. 5a). For the mouth-water experiment, there was a significant decrease in BET in the 100% treatment compared to the 0% control (p = 0.03) (Fig. 6b). There were also significant decreases for SLW/SLL in all treatments compared to the 0% control (p < 0.05 and p < 0.001) (Fig. 6a).

Figure 4.

The ratio of secondary lamella width (SLW) to secondary lamellar length (SLL) (a) and basal epithelium thickness (BET) (b) for the 21-d exposure period in the sodium chloride experiment. All data are reported as mean ± standard error. Statistical significance was analyzed using a one-way analysis of variance (ANOVA) providing assumptions of normality (Shapiro-Wilks) and homogeneity of variance (Levene's) were met. For data that did not meet these assumptions, the non parametric equivalent (Kruskal-Wallis test) was used. Differences were considered significant when p < 0.05. Statistical significance is denoted by * = p ≤ 0.05 versus 0% and ** = p ≤ 0.001 versus 0%.

Figure 5.

The ratio of secondary lamella width (SLW) to secondary lamellar length (SLL) (a) and basal epithelium thickness (BET) (b) for the 21-d exposure period in the sulphate experiment. All data are reported as mean ± standard error. Statistical significance was analyzed using a one-way analysis of variance (ANOVA) providing assumptions of normality (Shapiro-Wilks) and homogeneity of variance (Levene's) were met. For data that did not meet these assumptions the non parametric equivalent (Kruskal-Wallis test) was used. Differences were considered significant when p < 0.05. Statistical significance is denoted by * = p ≤ 0.05 versus 0% and ** = p ≤ 0.001 versus 0%.

Figure 6.

The ratio of secondary lamella width (SLW) to secondary lamellar length (SLL) (a) and basal epithelium thickness (BET) (b) for the 21-d exposure period in the mouth-water experiment. All data are reported as mean ± standard error. Statistical significance was analyzed using a one-way analysis of variance (ANOVA) providing assumptions of normality (Shapiro-Wilks) and homogeneity of variance (Levene's) were met. For data that did not meet these assumptions the non parametric equivalent (Kruskal-Wallis test) was used. Differences were considered to be significant when p < 0.05. Statistical significance is denoted by * = p ≤ 0.05 versus 0% and ** = p ≤ 0.001 versus 0%.

DISCUSSION

Water chemistry

Data obtained from various government and industry sources has been graphed, and significant trends in average concentrations of dissolved chloride, dissolved sodium, and sulphate both across the Athabasca River and over time (1966–2006) were computed 8. Using this data, we know that the highest average concentrations (across the years 1996–2006) of these parameters in the Athabasca River are 21.09 mg/L dissolved sodium, 22.71 mg/L dissolved chloride, and 45.27 mg/L sulphate. The concentrations of dissolved sodium and dissolved chloride in the 100% sodium chloride treatment (25.43 and 38.90 mg/L, respectively) and the concentration of sulphate in the 100% treatment of the sulphate experiment (101.37 mg/L) were chosen to imitate these levels (Table 1). Although the concentrations of dissolved sodium and chloride used in the present study were slightly higher than was seen in the Athabasca River by Squires et al. 8, they are still representative of the average concentrations of these parameters in the Athabasca River. The concentration of sulphate in the 100%, however, was approximately twice the amount of average sulphate measured in the Athabasca River. The concentration of sulphate in the 50% treatment (57.13 mg/L) was much closer to the average concentrations; therefore, these levels are still representative of the actually in-river concentrations of these parameters.

We also observed an unrepresentative spike in total organic carbon to 14.80 mg/L in the 12.5% treatment in the dissolved sodium experiment. However, because this is based on a single measurement, it is unlikely to be representative of the average conditions in this treatment during the entire experiment; therefore, it is not considered to be of particular concern. This is further realized when this level is outside of the averaged levels seen throughout all treatments in the mouth water experiment. There was an increase in TOC in the 100% treatment of the dissolved sulphate experiment to 6.90 mg/L. While this was also based on a single measurement, unlike the spike seen in the 12.5% treatment in the dissolved sodium, this is still within the averaged range seen in the mouth-water treatments and could therefore reasonably be considered representative of average conditions for this experiment.

Fish reproduction and gill histology

NaCl experiment. All treatments except the 12.5% treatment had higher cumulative egg production than the 0% control, although not statistically significant. This result is similar to what is seen in the mean/eggs/female/d endpoint, where all treatments had higher egg production than the control except for the 25% treatment. Low levels of salinity are often added in aquaculture practices to help reduce fish stress (due to handling, crowding, etc.) 22. It is known that the toxicity of other freshwater constituents (such as ammonia and nitrate) depends on salinity levels, decreasing with higher salinity 23. The induction of anti-stress reactions in algae, amphipods, and fish includes the induction of certain stress proteins, which can induce multiple stress resistance 24. At the lowest concentrations used in the present study (6.25%), it is possible that we are seeing this protective effect.

In the sodium chloride experiment, the only negative percentage of change from control for total eggs/female/day was found in the 6.25 (3.67 mg/L) and 12.5% (3.23 mg/L) treatments. This result differs from the cumulative eggs/female and the mean eggs/female/day endpoints, where increases in egg production occurred at these levels. Total eggs/female/day was calculated by totaling the number of eggs each replicate breeding trio produced across the entire exposure period, then dividing this by the number of days of exposure (21) and the number of females per treatment (two per replicate, three replicates per treatment, equalling six females per treatment). Because this is a total number and is based on 21 d of exposure, we cannot necessarily extrapolate this value past the 21 d of this specific experiment. The cumulative and mean endpoints are not static totals and instead illustrate the general trend of reproductive output during the experiment. This is considered more realistic of what can be expected in a natural population and can predict the potential population-level effects of these constituents more effectively.

Gills are the major site of respiration and are in close, constant contact with the water. As such, they are particularly vulnerable to contaminants such as salinity 25. In similarly conducted experiments in the laboratory using dissolved sodium, the SLW/SLL was found to be significantly higher than the 0% control in the 6.25% (30.33 mg/L) and the 100% (88 mg/L) treatments 15. In these same experiments, the greatest decreases in the percentage of change from control for total eggs/female/d in the laboratory dissolved sodium experiment was also in the 6.25 and 100% treatments. These results demonstrate a possible link between increasing gill diffusion distance and decreasing reproductive output in the fathead minnow.

In the present study, there was a significant increase in SLW/SLL in the 12.5% (3.23 mg/L Na) treatment (p < 0.001), which may have contributed to a non significant decrease in reproductive ability as measured by cumulative egg production. If the duration of the exposure period in this experiment had been increased past 21 d, it is possible that the decrease in cumulative egg production in the 12.5% treatment would become significant and reflect the changes seen in the gill epithelium as has been demonstrated by laboratory studies 15.

The greatest increases in reproductive output (cumulative eggs/female, mean eggs/female/day, and percentage of change from control for total eggs/female/day) in the sodium chloride experiment occurred in the highest treatments (50 and 100%). At these higher concentrations, it is possible that we are seeing a different type of salinity effect. A study conducted by Cataldi et al. 26 tested whether the presence of certain levels of salinity (freshwater, iso-osmotic water, and seawater) would affect the osmolarity capabilities of fish after being exposed to stress. They found that fish held in freshwater had a significant decrease in serum osmolality, whereas fish held in seawater had a significant increase in serum osmolality.

In the present study, the 100% treatment had a significantly larger diffusion distance (SLW/SLL). Although not significant, however, the 100% treatment also had a greater cumulative egg production than the control and a positive change from control for total eggs/female/day. In this treatment, it is possible that despite the larger diffusion distance, the increase in osmolality, which occurs with exposure to higher levels of salinity over a longer period of time (such as the 21-d exposure period used in the present study), could have reduced stress, allowing fish to allot more energy to reproduction.

Additionally, the dissolved chloride concentration in the 100% treatment was 38.90 mg/L. In previous laboratory studies 15, the ideal amount of dissolved chloride for fish reproduction was found to be between 22.22 and 49.56 mg/L. The only treatment to fall within this range in the present experiment was the 100% treatment. It is possible that the effects of the presence of this higher level of dissolved chloride (through competitive binding) would offset the effects of dissolved sodium, allowing for an increase in reproduction ion this treatment.

Mouth-water experiment

The presence of turbidity and/or organic carbon in water collected from the main stem of the Athabasca River downstream of the oil sands operation and its potential effect on fathead minnow reproduction has not been studied previously. There was an increase of 175% in the 50% treatment (8.20 mg/L TOC). The highest level of mean egg production was significantly greater (p < 0.001) than the 0% control treatment and occurred in the 50% mouth-water treatment. The 50% mouth-water treatment also had the highest level of turbidity at 5.58 NTU. Turbidity consists of many aspects including sediment, algal cells, dissolved humic substances, dissolved minerals, and detrital organic matter 27. Humic substances are very complex organic molecules that can make up to most (80%) of the dissolved organic matter present in freshwater ecosystems 28. The presence of humic substances can evoke anti-stress reactions in exposed organisms in their effort to try to remove these easily accumulated substances 24. Low levels of humic substances have been shown to induce phase I and II enzymes, providing a protective effect; in the case of some amphipods, it also increases the number of offspring produced 24. Therefore, exposure to low concentrations of humic substances can help train the defense system and lead to stress resistance, thereby allowing for more energy to be diverted to other activities such as reproduction.

Located immediately upstream of the collection site for the mouth water used in the present study are several large oil sands mining and extraction sites. A study conducted by Tetreault et al. 29 sampled two small forage fish species (slimy sculpin and pearl dace) from several sites along the Athabasca River near the oil sands operations and compared results with fish sampled upstream of the oil sands operations. No significant differences in length, weight, condition factor, LSI, or GSI were found in either species. However, some differences in levels of steroid production at sites downstream of the oil sands operations were demonstrated. A more recent study by Kavanagh et al. 5 assessed the potential of aged OSPW on fathead minnow reproduction. These authors demonstrated a significant decrease in cumulative number of eggs for fathead minnows exposed to aged OSPW compared to reference water. The Kavanagh study also demonstrated a decrease in plasma steroid levels (testosterone, 11-ketotestosterone, and 17β-estradiol) in some cases. These studies focused on the impact of aged OSPW, which is currently held on site and not released into the Athabasca River adjacent to the oil sands operations.

Presently, no study has yet assessed the impact of Athabasca River water downstream of the oil sands operations on the reproductive potential of the fathead minnow. In the present study, a negative percentage of change from control for total eggs/female/d was seen in the 25% treatment, and a positive percentage of change from control was seen in the 50% treatment. It is unclear from the present study what factors might be attributed to these different changes. However, due to the complex nature of the oil sands process water and the natural abundance of oil sands in the surrounding area, it is important to clarify the potential impacts of actual in-river concentrations of oil sands related water quality parameters on the reproduction of local fish species.

A study conducted using aged process-affected water from the oil sands operations near the mouth of the Athabasca River showed that although the water was not acutely toxic to yellow perch or goldfish, there were impacts indicated on gills 30. These included changes in mucous cell proliferation; the consequence of these changes was an increase in the distance for gas exchange along the secondary lamellae, potentially reducing the efficiency of gas exchange. In contrast, all treatments in the present study had significantly lower diffusion distances (SLW/SLL) than the 0% control. This result would imply that increasing amounts of mouth water have a positive effect (lower diffusion distance) on the gill epithelium. One reason for this could be that the ionic potential of the mouth water is closer to equilibrium with the fish versus the ionic potential of the headwater (0% control). A greater difference in electric potential causes increase stress and damage to the animal due to the need to allocate greater energy to osmoregulation 31, 27. This is further reflected in the cumulative egg production, where all treatments (except the 25%) had higher cumulative egg production than the 0% control. This trend is also similar to what was observed in the laboratory experiments conducted with dissolved sodium, in which a link between decreased diffusion distance and increased reproductive output was demonstrated 15. Because the trends in cumulative egg production in the present study mirrored those seen in this previous laboratory work, it can be reasonably assumed that if the duration of exposure was increased, these trends may become significant.

The primary role of the gill epithelium in aquatic organisms is to support life through gas exchange, acid-base regulation, nitrogenous waste excretion, immunity, and ion transport 31. The fish gills are considered one of the most sensitive organs to external pollution exposure due to their direct contact with the water. As such, several methods have been developed to assess morphological and physical changes to this area 22. Previous research has shown a link between decreasing diffusion distance (SLW/SLL) and increasing reproductive output and has demonstrated that changes in the gill functions can impact the reproductive performance of aquatic biota 15.

There were higher levels of TOC in the 12.5 (4.27 mg/L), 25 (4.80 mg/L), 50 (8.20 mg/L), and 100% (10.73 mg/L) treatments, which can also contribute to the trends in gill damage and consequently reproductive output. Organic matter has been shown to influence the flux of Na+ across gills of freshwater organisms such as Daphnia and fish 31. In past decades, growing recognition that organic matter can affect the physiology of organisms through several mechanisms including activation of glutathione S-transferase, induction of heat shock proteins, and CYP1A enzymes, as well as changes in behavior 31. The presence of organic matter at concentrations of 10 mg/L can also alter the fundamental physiological properties of fish gills by hyperpolarizing gill membranes 31. This happens when humic substances complex with biologically active ions, such as free Ca2 + , from the water. Free Ca2+ is an important constituent of epithelial tight junctions, and reductions in this ion to very low levels are known to cause a general increase in diffusive ion losses, as well as selectively enhance the permeability of the gills to Na+ relative to Cl–, resulting in hyperpolarization 31, 32.

Sulphate experiment

There was no difference among treatments for cumulative eggs/female, total eggs/female/d and mean eggs/female/d in the dissolved sulphate experiment compared to the 0% control. However, at the end of the experiment (day 21 of exposure), all treatments had lower cumulative egg production than the 0% control. Although not significant, there was also an increase in mean egg production in the 25% treatment in the dissolved sulphate experiment.

There have been a few studies on the toxicity of sulphate in the aquatic environment. The water flea, Hyalella azteca, was found to have an LC50 of 512 mg/L SO4 33. Toxicity associated with excess SO4 can be related to indirect effects on calcium availability rather than direct impacts from SO4 11. The toxicity of SO4 to aquatic biota was found to decrease with increasing hardness and chloride concentrations 33. The presence of calcium and chloride can aid in the biota's ability to osmoregulate and therefore tolerate higher ionic solutions 33, 11.

The province of British Columbia, Canada has established a water quality guideline to protect aquatic life for sulphate of 100 mg/L 34. In the dissolved sulphate experiment, three treatments had lower percentage of change from control for total eggs/female/d than the control, that is, 6.25 (18.57 mg/L), 12.5 (23.33 mg/L) and 100% (101.37 mg/L). The highest of these treatments had a concentration comparable to this guideline set by British Columbia. This guideline was based on the acute toxicity data generated from several species of invertebrates, fish, algae, moss and amphibians with a twofold safety margin. Our results are based on chronic exposures (21 d) to sublethal levels of sulphate to assess the impacts on the reproductive output on fathead minnows. Because this water quality guideline is based on short-term exposures (<24 h) on lower trophic-level species (C. dubia), it is unlikely that our lowest concentration in which an effect is observed would correlate with this guideline value. It would therefore be worthwhile to conduct further studies on the sublethal effects of sulphate on the population of higher order species.

In the dissolved sulphate experiment, there was a significant increase in BET in the 50% treatment (p = 0.039) and significant decrease in SLW/SLL in the 6.25, 12.5, and 100% treatments (p < 0.05) compared with the 0% control. Both BET and SLW/SLL are measures of gill epithelium thickness. Decreases in either of these measurements correspond to decreases in the diffusion distance across the gill. These treatments also all had a negative change from the control for total eggs/female/d. Therefore, we saw decreased gill diffusion distances in the treatments that had the lowest cumulative egg production. This is opposite from what was observed in the sodium chloride experiment in the present study, as well as was observed in previous laboratory studies 15.

One explanation for these conflicting results is the presence of higher turbidity in the 6.25 (5.13 NTU) and 12.5% (5.07 NTU) treatments and higher TOC (6.90 mg/L) in the 100% treatment relative to the rest of the treatments in the present study (Table 1). While we saw increased reproduction of fathead minnow likely in response to increased levels of TOC and turbidity in the mouth-water experiment, there are thresholds beyond which adverse effects on fish populations are observed. It is possible that the presence of higher levels of organic matter in these treatments can have a negative effect on fish reproduction.

The reproductive success of salmonid fish are known to be especially sensitive to suspended organic matter since matter depositing on the stream beds will block pores in the gravel, which are ideal for depositing eggs 27. In general, though, high enough levels of suspended organic matter are known to clog fish gills, which decreases the ability of the fish for oxygen exchange and osmoregulation 35. Also, suspended sediments can cause stress by suppressing the immune system, which can then lead to susceptibility to disease 27. Therefore, sublethal thresholds that account for the effects of turbidity on reproductive potential are necessary and important when considering the impacts of contaminants on the aquatic ecosystem.

CONCLUSIONS

Previous research has noted that dissolved sodium, chloride, and sulphate concentrations have significantly increased in the lower reaches of the Athabasca River over 30 years 8. Although laboratory research has shown that both sodium and chloride have been shown to alter reproduction in fathead minnow and gill membrane diffusion distances, it was unknown whether these effects would be seen using actual in-river water concentrations 15.

The present study aimed to determine whether the thresholds developed in the laboratory could be applied using water taken from both the headwaters and mouth of the Athabasca River. This was done by exposing fish to concentrations of sodium chloride and sulphate relevant to average levels in the mouth of the river. In addition, water taken from downstream of industrial inputs in the mouth of the river was used for threshold development. In the dissolved sulphate experiment, the treatments that had the lowest reproductive output were the 6.25 (18.57 mg/L) and the 12.5% (23.33 mg/L) treatments. Current guidelines also state that dissolved sulphate should not exceed 100 mg/L 33. Based on this information, the ideal amount of dissolved sulphate for fathead minnow reproduction was determined to between 23.33 and 100 mg/L. The greatest increases in reproductive output (cumulative eggs/female, mean eggs/female/d, and percentage of change from control for total eggs/female/d) in the sodium chloride experiment occurred in the highest treatment (25.43 mg/L Na and 38.90 mg/L Cl). This is very similar to the range identified in the laboratory study for dissolved sodium of between 36.11 and 57 mg/L and dissolved chloride between 22.22 and 49.56 mg/L 15.

Although differences in cumulative egg production in the present study were not found to be significant in the sodium chloride experiment, the trends between gill diffusion distances and cumulative reproduction are similar to those demonstrated in previously conducted laboratory studies using higher concentrations of dissolved sodium and chloride 15. The present study used concentrations in which the 100% treatment was 30% of the concentrations used in the 100% treatment in the laboratory studies. It is unlikely that at these lower levels we would see the same significant effects on reproduction as were seen in the higher concentrated laboratory studies over the similar 21-d exposure period. However, because we did see a similar connection between gill diffusion distance and reproduction output, as was demonstrated in the laboratory study, this may over a longer period of time have a more significant impact on reproductive output.

It is important to determine thresholds that are specific to an area or region. Further research would include longer exposure periods to determine if the trends seen in the present study continue and influence future generations of fathead minnow. It would also be important to use multiple species, because there are differing sensitivities among species to these types of exposures. This information would allow us to form a better understanding of how these parameters will affect this river basin in the long-term.

Acknowledgements

Funding for the present study was obtained through the Canada Research Chairs Program, Canadian Foundation for Innovation and a National Sciences and Engineering Research Council Discovery Grant. The authors acknowledge M. Heggstrom, S. Pryce, D. Duro, and M. Driessnack for their technical assistance during these experiments. We also acknowledge P. Jones and J. Raine for their assistance with gill histology and interpretation. Dilution systems were built by J. Mollison of JCM specialties, Saskatoon, SK. We gratefully acknowledge the logistical support of L. Zaffino and R. Dierker of ATCO Electric for providing us with the electrical power and space required to run the trailer. We also thank W. Hughson, warden at Jasper National Park for support and assistance in providing us with the permits to perform these experiments in a national park.

Ancillary