Non-point source discharges, such as agricultural runoff, are often complex mixtures of chemical and non-chemical stressors. The complexity of runoff is compounded by its sporadic releases and few studies have attempted to assess the impacts of runoff on aquatic biota. In this study, an effects based approach was used to examine survival and reproduction of slimy sculpin (Cottus cognatus) in the intensive potato-farming areas of northwestern New Brunswick, Canada. Using non-lethal methods, fish were collected during the ice-free months through a gradient of agricultural intensity. These data were correlated with waterborne levels of pesticides, water temperatures and precipitation data. Results indicate that both adult and young-of-the-year (YOY) fish are longer and heavier in the downstream sites draining areas of higher agricultural intensity. Precipitation has a significant negative relationship with %YOY in the agricultural areas but not in the upstream forested area, indicating that contaminants are present in runoff caused by intensive rainfall events. Our results indicate that YOY sculpin may be at higher risk in the agricultural areas in years of heavier summer rains where peaks in pesticide levels occur. This study expands the existing knowledge base and development of non-lethal methods to define cause–effect relationships.

INTRODUCTION

Non-point source pollution can negatively impact aquatic biota by altering physical habitat, seasonal water flow, systemic food base, water quality, and by modifying interactions among organisms (Karr 1999; Potter et al. 2004). Resident biota of aquatic ecosystems can serve as continuous monitors of the cumulative effects of these multiple stressors on those systems, and are often considered in environmental assessments (Munkittrick et al. 2000; Diamond & Serveiss 2001). Research on the impacts of agriculture on ecosystem health has primarily focused on single measures or stressors such as soil erosion (Chow & Rees 1995; Pimental et al. 1995), pesticide use (Culliney et al. 1992; Clark et al. 1999; Battaglin & Fairchild 2002), or contamination of groundwater by fertilizer and other agricultural chemicals (Bouwer 1990; Napier & Brown 1993; Böhlke 2002). While these studies represent important contributions to understanding the impacts on ecosystem health, information on single stressors provides only a partial picture of the integrated impacts of agricultural stressors. Cumulative effects on ecosystem health are better assessed by characterizing the integrated responses of upper ecosystem-level indicators (Munkittrick et al. 2000).

There have been a number of studies on agricultural activities along the Little River watershed, located north of Grand Falls, New Brunswick (NB), Canada (Figure 1). The Little River is a fourth order stream that originates in a forested landscape and drains predominantly agricultural lands. It is a tributary of the St John River. The St John watershed from Grand Falls to Hartland is one of the largest potato-farming regions in eastern Canada. The Black Brook watershed is a 1,450 ha sub-basin within the Little River watershed, dominated by agricultural land use (66%) with potatoes as the most dominant crop (Figure 1). It represents one of the most intensely farmed watersheds in eastern Canada and is routinely monitored by Agriculture and Agri-Food Canada for crop management and improvement (Chow et al. 2000; Rees et al. 2002).

Figure 1

Map of the Little River System showing the Black Brook sub-basin. Non-lethal sampling sites are consistent with the studies of Gray et al. (2005). Weir 1 site has permanent gauging and sampling station for pesticide analysis.

Figure 1

Map of the Little River System showing the Black Brook sub-basin. Non-lethal sampling sites are consistent with the studies of Gray et al. (2005). Weir 1 site has permanent gauging and sampling station for pesticide analysis.

Potato production in northwestern NB typically consists of a potato-grain or potato-potato-grain rotation with intensive soil manipulation and heavy use of chemicals. Rees et al. (2007) described typical farming practices for grain and potato production. Given the continental-like weather (high-intensity summer thunder storms), shallow soil (developed in basal till materials), rolling landscapes (slopes commonly over 5%), and the limited soil protection (potato crop rotations with bare soil in between), soil and water conservation practices are required to preserve the land. Regardless, not all lands are adequately protected. Significant acreages are cultivated up and down slope with serious consequences. Numerous studies of soil loss under potato cropping have been reported for NB and adjacent Maine, USA. Given the soil and landscape conditions in northwestern NB and the cultural practices and considerable time the soil is bare of cover under potato-grain rotations, soil erosion is a major problem as it drains into the adjacent aquatic environment and has the potential to impact the resident fish community.

Fish community structure is limited in northern NB with a maximum of three species present at most sites: brook char (Salvelinus fontinalis), slimy sculpin (Cottus cognatus), and brook stickleback (Culea inconstans) (Curry & Munkittrick 2005). Sculpin populations were selected for monitoring based on their use in previous studies in this watershed. Recent studies have documented impacts on slimy sculpin populations at multiple locations along a gradient within this watershed (Gray et al. 2005), in a comparative study of multiple rivers with agricultural gradients (Gray & Munkittrick 2005) and at single locations within 21 different watersheds (Welch et al. 1977; Gray et al. 2005). The impacts included increases in growth and condition factor; decreases in liver and gonad size, fecundity, nest abundance, and nest size; and decreases in densities of young-of-the-year (YOY) and adult sculpin. Growth and mortality are important population-level dynamics influencing the ecology of fish populations by directly influencing the role of individuals within the community and interactions among species, especially in size-structured populations (Werner & Gilliam 1984). This latter study also documented numerous stressors that may be contributing to the fish response patterns, including changes in temperature, nutrients, and sedimentation (Gray & Munkittrick 2005) but did not determine the relative importance of the various stressors.

The objective of the present study was to monitor fish populations in an effort to characterize the fish responses. Temporal collections will identify the timing of potential mortality events and investigate reductions in fish numbers previously documented in agricultural areas. Non-lethal sampling approaches (as in Gray et al. 2002) were chosen due to repeat sampling over multiple sites, and the limited mobility of sculpin in these areas (Gray et al. 2004; Cunjak et al. 2005). To examine the influences of other stressors, data were also collected regarding precipitation and storm intensity, temperature, and in-stream pesticide levels. Efforts were focused on expanding the existing knowledge base and the development of methods to define cause–effect relationships.

METHODS

Sculpin collections

Previous work in this watershed (Gray et al. 2005) provided a basis for selecting sites along the gradient of potato cultivation intensity. The main study sites for fish sampling in this project were along an increasing gradient of potato-farming intensity in the Little River watershed (downstream of Ten Mile Brook at 47 °09′95″N 67 °40′10″W, draining areas with no agricultural activity; downstream of Donat Brook at 47 °05′95″N 67 °42′00″W, draining areas of intermediate agricultural intensity; downstream of Dead Brook at 47 °04′85″N 67 °42′95″W, draining areas of more intensive agricultural activity), located north of Grand Falls, NB, Canada (Figure 1). Watershed area and land use within each sub-watershed are listed to illustrate the agricultural gradient (Table 1). Values are shown as total area in hectares as well as percentage land use within the watershed.

Table 1

Values for watershed area and land use shown in both total area in hectares and percentage of total watershed

Watershed Total area (ha) Agricultural land (ha) Forestry (ha) Others (ha) Agriculture (%) Forest (%) Other (%) 
Ten Mile Brook 1560.23 334.11 1074.84 151.28 21.41 68.89 9.70 
Dead Brook 6688.18 3206.75 3025.71 455.72 47.95 45.24 6.81 
Donat Brook 1455.12 1044.16 344.63 66.33 71.76 23.68 4.56 
Watershed Total area (ha) Agricultural land (ha) Forestry (ha) Others (ha) Agriculture (%) Forest (%) Other (%) 
Ten Mile Brook 1560.23 334.11 1074.84 151.28 21.41 68.89 9.70 
Dead Brook 6688.18 3206.75 3025.71 455.72 47.95 45.24 6.81 
Donat Brook 1455.12 1044.16 344.63 66.33 71.76 23.68 4.56 

Slimy sculpin were collected by sampling shallow (approximately 0.2–0.75 m depth), faster runs, and riffles (approximately 1.1–1.5 m/s) with boulder/cobble substrates using a backpack electrofisher (Smith-Root type VII) and dip nets (6 mm mesh size). Collections in 2002 targeted the first 100 sculpin collected. Upon review of the 2002 data, subsequent collections were adapted to continue until a minimum of 100 adult fish were caught to increase the resolution of size frequencies in the age classes after YOY emergence. All YOY were collected and measured. Barrier nets were not installed, as we have previously found no significant differences with sculpin collection in open versus closed sites using one sweep through an area (Gray et al. 2002). Non-lethal sampling of all fish involved species identification and measurements of fork length, or total length for sculpin (±1 mm), and weight (±0.01 g). All fish were then released back into the site where they were collected. Size frequency data were used to examine age distributions and condition factors for the fish (Gray et al. 2002). Condition factor was calculated as (weight/length3)*100,000, with length reported in millimeters. Normality and homoscedasticity were assessed by visual examination of normal probability and residual plots, respectively. YOY were discriminated by plotting length–frequency distributions for each site. The relationships between mean daily temperature fluctuation and total rainfall, and YOY body size and abundance were assessed using linear regression. Statistical analyses were completed using Systat© (v. 9, SPSS, Chicago, IL, USA). Length–weight relationships were analyzed using the ANCOVA (analysis of covariance) function in Systat©.

Temperature data

Temperature recorders (12-bit, Minilog-TR, Vemco Limited, Shad Bay, NS, Canada) were placed at each site to record hourly water temperatures. Temperature was recorded beginning in May following spring runoff until mid-October to encompass the period of potential growth for YOY sculpin, from the time of approximate emergence from the nests to the end of growth for the first growing season. Degree-days (above 0 °C) were calculated as sum of daily temperature between 27 July and 18 October.

Pesticide analyses

All solvents used were of distilled in glass grade from Caledon (Georgetown, ON, Canada). Water used was of high performance liquid chromatography (HPLC) grade from Fisher Scientific (Nepean, ON, Canada). All pesticide standards used were of Pestanal grade from Supelco-Sigma-Aldrich (Oakville, ON, Canada). [13C3]-Atrazine and [13C6]-Carbaryl were from Cambridge Isotope Laboratories, Inc. (Andover, MA, USA). All HPLC solvents were filtered prior using 0.2 μm membrane filters from Chromatographic Specialties (Brockville, ON, Canada). Five percent dimethyldichlorosilane in toluene from Supelco-Sigma-Aldrich was used in the solid phase extraction (SPE) manifold from Supelco-Sigma-Aldrich.

The pesticides analyzed were comprised of six compound classes (nicotinyl, carbamate, organophosphate, triazinone, urea, and phenylamine) and some of the major in-use pesticides associated with potato production in the Black Brook watershed (imidacloprid, linuron, metribuzin – based on annual farmer land use surveys conducted prior to our study period). Pesticides included three insecticides, carbaryl, azinphos-methyl, and imidacloprid; three herbicides, linuron, metobromuron, and metribuzin; and one fungicide, metalaxyl. Surface water samples (1 L) were collected at the outlet of Black Brook at a permanent gauging and sampling station (Weir 1; Figure 1). This site consisted of an instrument shelter housing a stilling well with a below-ground connection to the stream channel for stage height measurements. From which, flow rates were calculated corresponding to the date and time of sample collections. Weir 1 had an automated Isco® system located in-stream for water sample collections which were collected weekly at a predetermined time. The sampler also collected samples with increased intensity to follow hydrographic peaks during individual rain events. A non-absorbing nylon tubing was used between the stream water sampler intake (located at a site of maximum mixing) and the automated Isco® water sampler and the water samples were collected in glass bottles. Samples were transported to the laboratory and stored at 4 °C until extraction, normally within 2 days.

Prior to extraction, all samples were filtered through 0.2 μm glass fiber filters (GF/C; Whatman, Fisher Canada, Oshawa ON, Canada). Filtered samples were spiked with the surrogate [13C6]-Carbaryl and extracted using a SPE procedure that employed reversible graphitized carbon black (GCB) SPE cartridges (80–100 mesh, 500 mg, 2 mL reversible; Supelco-Sigma-Aldrich). The extraction method was modified from Di Corcia et al. (1993) to avoid sample preservation and stability problems typical of pesticide analysis. The GCB solid phase was also selected as it offers the best all-round recovery of multi-residue analysis including polar compounds, with the added benefit of preserving the analytes on the cartridges following extraction (Hennion 2000). GCB cartridges were placed on a SPE manifold (Supelco-Sigma-Aldrich) and conditioned with 10 mL of 80:20 DCM:MeOH, 4 mL of methanol, then with 10 mL of HPLC grade water. All conditioning steps were carried out using gravity elution. Elution times varied; with flow rates of approximately 15 mL/min, each sample took approximately 60 minutes to extract. Once all the water had passed through, samples were then passed through the cartridges at 15 mmHg vacuum for an additional 2 minutes as a drying step. Cartridges were then sealed into polypropylene bags and frozen at −20 °C for shipment to Burlington, ON for elution and analysis by liquid chromatography tandem mass spectrometry (LC–MS–MS). After arriving in Burlington, cartridges were removed from the freezer and thawed to room temperature. In order to remove any excess water contained in the GCB bed, 800 μL of MeOH was placed onto the cartridge and gravity eluted, followed by 15 mm Hg vacuum for 2 minutes. Cartridges were then reversed on the SPE manifold and eluted by gravity with 10 mL of 80:20 DCM:MeOH into 15 mL graduated centrifuge tubes. Eluates were evaporated under a gentle stream of nitrogen. Residues were then redissolved in 50 μL [13C3]-Atrazine instrument performance standard and analyzed by LC–MS–MS.

Pesticide separations were achieved with an Agilent 1100 liquid chromatograph fitted with a C18 3.5 μm, 2.1 × 100 mm column (Waters Symmetry, Brampton, ON, Canada) and eluted with 50/50 acetonitrile/water for 6 minutes isocratic at a flow rate of 300 μL/min. The chromatograph was fitted to a triple quadrupole tandem mass spectrometer (MDS Sciex API2000, Toronto, ON, Canada) configured with an atmospheric pressure photoionization (APPI; Photospray, MDS Sciex) ion source receiving 30 μL/min toluene dopant. All mass spectrometry conditions were optimized for each pesticide by infusion, and quantitations were performed using multiple reaction monitoring mode for the most abundant transition unique to each compound. Negative ion APPI was used for metobromuron and linuron, while positive ion transitions were used to quantify all other compounds. Quantitations were made using external calibrations standards acquired after every 20 sample injections. Method detection limits ranged from 0.05 to 10.4 ng/L.

Precipitation data

Precipitation data were reported as monthly and annual total rainfall (mm) at the St Leonard airport (Station 6256, Canada National Climate Archive). Storm intensity was characterized as >15 mm total rainfall. Storms must be of a certain duration and intensity to mobilize sediments and potentially pesticides off the fields and into the nearby watercourses (Waters 1995). Rainfall amount was used in our study as an indicator of erosivity.

RESULTS

Sculpin population assessments

Fish population data were analyzed monthly for each site by generating frequency histograms based on total length. Statistical analysis options for these data were limited, and these data were assessed visually to determine YOY size and growth and summarized (Table 2). At all three sites, the YOY size class was easily distinguished, appearing in August, and increasing in size over time, as described previously (Gray et al. 2002). At the agricultural site (Dead), August YOY sculpin were significantly longer than both upstream sites in all 3 years, but had a similar condition compared to reference (Ten Mile) site fish (Table 2). The proportion of the population composed of YOY increased at the upstream reference site (Ten Mile) between August and September, but remained stable during the rest of the fall period (Oct) in both 2002 and 2004; downstream (agricultural) YOY made up a larger portion of the population in 2002 and 2003, but not in 2004 (Table 2). There were no consistent differences in condition or fish size early in the year (Table 3). Between July and October, adult fish mean length and weight increased 8% and 17%, respectively, at the Ten Mile site, 10 and 20% at the Donat site, and 13 and 38% at the Dead site (Table 3).

Table 2

Young of the year (YOY) slimy sculpin collected from monthly monitoring of Little River along a gradient of agricultural inputs (2002–2004)

  Ten Mile (forested) Donat (intermediate) Dead (agriculture) 
n POP Size (mm) Body Wt (g) Length K n POP Size (mm) Body Wt (g) Length K n POP Size (mm) Body Wt (g) Length K 
2002 
 August 21 0.21 25 0.16 25.10 1.01 37 0.38 29 0.24 29.19 0.93 71 0.72 30 0.17 30.31 1.01 
 September 47 0.44 31 0.39 32.17 1.15 59 0.61 32 0.37 32.05 1.11 80 0.80 34 0.46 33.80 1.16 
 October 45 0.45 31 0.38 31.78 1.14 47 0.47 36 0.49 36.09 1.03 69 0.69 36 0.56 37.03 1.08 
 November 47 0.46 34 0.39 33.77 0.97 44 0.61 36 0.53 36.50 1.07        
2003 
 August 47 0.29 27 0.18 26.72 0.94 21 0.18 25 0.16 25.33 0.97 80 0.45 27 0.26 30.06 0.92 
 October 53 0.34 32 0.35 32.25 1.02 39 0.31 32 0.34 30.92 1.16 88 0.50 36 0.47 36.07 0.98 
2004 
 August 65 0.43 26 0.14 23.33 1.08 20 0.30 27 0.19 26.42 1.01 12 0.24 28 0.14 27.36 1.07 
 September 72 0.52 31 0.35 31.59 1.11 72 0.53 32 0.35 31.90 1.05 119 0.59 32 0.36 32.26 1.06 
 October 100 0.56 35 0.45 34.79 1.05 65 0.40 35 0.45 35.61 0.98 74 0.44 35 0.44 35.27 0.96 
  Ten Mile (forested) Donat (intermediate) Dead (agriculture) 
n POP Size (mm) Body Wt (g) Length K n POP Size (mm) Body Wt (g) Length K n POP Size (mm) Body Wt (g) Length K 
2002 
 August 21 0.21 25 0.16 25.10 1.01 37 0.38 29 0.24 29.19 0.93 71 0.72 30 0.17 30.31 1.01 
 September 47 0.44 31 0.39 32.17 1.15 59 0.61 32 0.37 32.05 1.11 80 0.80 34 0.46 33.80 1.16 
 October 45 0.45 31 0.38 31.78 1.14 47 0.47 36 0.49 36.09 1.03 69 0.69 36 0.56 37.03 1.08 
 November 47 0.46 34 0.39 33.77 0.97 44 0.61 36 0.53 36.50 1.07        
2003 
 August 47 0.29 27 0.18 26.72 0.94 21 0.18 25 0.16 25.33 0.97 80 0.45 27 0.26 30.06 0.92 
 October 53 0.34 32 0.35 32.25 1.02 39 0.31 32 0.34 30.92 1.16 88 0.50 36 0.47 36.07 0.98 
2004 
 August 65 0.43 26 0.14 23.33 1.08 20 0.30 27 0.19 26.42 1.01 12 0.24 28 0.14 27.36 1.07 
 September 72 0.52 31 0.35 31.59 1.11 72 0.53 32 0.35 31.90 1.05 119 0.59 32 0.36 32.26 1.06 
 October 100 0.56 35 0.45 34.79 1.05 65 0.40 35 0.45 35.61 0.98 74 0.44 35 0.44 35.27 0.96 

YOY were not present in nets until August of all years and sampling was conducted in ice-free months only at a constant catch per unit effort of 10,000 electrofishing seconds. All fish were measured and released back into collection areas. POP refers to the proportion of the population sample collected represented by YOY. Size of YOY is reported as median size of the sample. K is the condition factor calculated as (weight/length3)*100,000, with weight reported in grams and length is reported in millimeters.

Table 3

Data collected from monthly sampling of adult slimy sculpin (2002–2004)

  Ten Mile (forested)
 
Donat (intermediate)
 
Dead (agriculture)
 
n Length (mm) Body wt (g) K n Length (mm) Body wt (g) K n Length (mm) Body wt (g) K 
2002 
 July 99 60.13 (1.02) 2.49 (0.14) 1.19 (0.03)a 100 57.54 (0.94) 2.48 (0.12) 1.23 (0.02)a 100 59.45 (1.01) 2.49 (0.14) 1.09 (0.02)b 
 August 77 56.70 (0.99) 1.90 (0.11) 0.99 (0.03) 63 59.06 (0.99) 2.17 (0.13 0.99 (0.03) 29 61.52 (1.23) 2.50 (0.16) 1.04 (0.01) 
 Septembera 58 60.41(1.27) 2.29 (0.15) 1.09 (0.02) 38 57.42 (1.26) 2.23 (0.18) 1.10 (0.01) 20 61.60 (1.50) 2.80 (0.23) 1.15 (0.02) 
 October 54 58.57 (1.43) 2.29 (0.22) 1.01 (0.01) 54 61.17 (1.00) 2.53 (0.14) 1.05 (0.01) 31 64.45 (1.21) 3.07 (0.17) 1.11 (0.02) 
 November 53 60.41 (1.34) 2.43 (0.20) 1.00 (0.01) 28 62.21 (1.41) 2.58 (0.18) 1.03 (0.02)     
2003 
 June 101 49.18 (1.41) 1.68 (0.17) 1.10 (0.02)a 101 48.64 (1.23) 1.58 (0.14 1.14 (0.01)a 99 52.08 (1.15) 1.88 (0.14) 1.18 (0.01)b 
 Julya 110 54.21 (1.07) 1.80 (0.11) 0.99 (0.01) 101 52.71 (1.06) 1.49 (0.10) 0.88 (0.01) 105 58.35 (1.25) 2.24 (0.19) 0.94 (0.01) 
 September 112 57.52 (0.97) 2.00 (0.12) 0.94 (0.01)a 98 58.18 (1.06) 2.01 (0.12) 0.92 (0.01)b 99 63.99 (1.03) 2.65 (0.15) 0.93 (0.01)a 
 October 101 58.79 (0.88) 2.20 (0.12) 1.01 (0.01)a 88 58.79 (0.95) 2.19 (0.12) 1.01 (0.01)a 89 65.58 (0.99) 3.25 (0.17) 1.04 (0.01)b 
2004 
 May 100 40.62 (1.05) 0.83 (0.04) 0.99 (0.02)a 97 48.53 (0.95) 1.49 (0.12) 1.10 (0.02)b 99 43.15 (1.27) 1.06 (0.15) 0.99 (0.02)a 
 June 51 50.04 (1.73) 1.93 (0.23) 1.12 (0.02)a 38 55.42 (1.97) 2.11 (0.21) 1.10 (0.03)a 47 55.23 (2.11) 2.18 (0.30) 1.04 (0.02)b 
 Julya 92 53.91 (1.03) 2.00 (0.13) 1.15 (0.02) 91 58.53 (1.41) 2.79 (0.18) 1.19 (0.03) 92 51.88 (0.82) 1.70 (0.07) 1.20 (0.03) 
 Augusta 87 57.09 (0.70) 1.98 (0.08) 1.02 (0.01) 47 62.06 (1.64) 2.80 (0.23) 1.06 (0.01) 39 62.08 (1.78) 2.69 (0.26) 1.07 (0.03) 
 September 57 62.26 (1.53) 2.86 (0.29) 1.03 (0.01) 98 62.09 (0.99) 2.70 (0.06) 1.04 (0.02) 103 59.24 (0.95) 2.30 (0.15) 1.01 (0.01) 
 October 77 62.39 (1.04) 2.57 (0.17) 0.98 (0.01)a 100 66.39 (1.04) 3.14 (0.17) 0.99 (0.01)a 97 62.05 (0.98) 2.47 (0.15) 0.95 (0.01)b 
  Ten Mile (forested)
 
Donat (intermediate)
 
Dead (agriculture)
 
n Length (mm) Body wt (g) K n Length (mm) Body wt (g) K n Length (mm) Body wt (g) K 
2002 
 July 99 60.13 (1.02) 2.49 (0.14) 1.19 (0.03)a 100 57.54 (0.94) 2.48 (0.12) 1.23 (0.02)a 100 59.45 (1.01) 2.49 (0.14) 1.09 (0.02)b 
 August 77 56.70 (0.99) 1.90 (0.11) 0.99 (0.03) 63 59.06 (0.99) 2.17 (0.13 0.99 (0.03) 29 61.52 (1.23) 2.50 (0.16) 1.04 (0.01) 
 Septembera 58 60.41(1.27) 2.29 (0.15) 1.09 (0.02) 38 57.42 (1.26) 2.23 (0.18) 1.10 (0.01) 20 61.60 (1.50) 2.80 (0.23) 1.15 (0.02) 
 October 54 58.57 (1.43) 2.29 (0.22) 1.01 (0.01) 54 61.17 (1.00) 2.53 (0.14) 1.05 (0.01) 31 64.45 (1.21) 3.07 (0.17) 1.11 (0.02) 
 November 53 60.41 (1.34) 2.43 (0.20) 1.00 (0.01) 28 62.21 (1.41) 2.58 (0.18) 1.03 (0.02)     
2003 
 June 101 49.18 (1.41) 1.68 (0.17) 1.10 (0.02)a 101 48.64 (1.23) 1.58 (0.14 1.14 (0.01)a 99 52.08 (1.15) 1.88 (0.14) 1.18 (0.01)b 
 Julya 110 54.21 (1.07) 1.80 (0.11) 0.99 (0.01) 101 52.71 (1.06) 1.49 (0.10) 0.88 (0.01) 105 58.35 (1.25) 2.24 (0.19) 0.94 (0.01) 
 September 112 57.52 (0.97) 2.00 (0.12) 0.94 (0.01)a 98 58.18 (1.06) 2.01 (0.12) 0.92 (0.01)b 99 63.99 (1.03) 2.65 (0.15) 0.93 (0.01)a 
 October 101 58.79 (0.88) 2.20 (0.12) 1.01 (0.01)a 88 58.79 (0.95) 2.19 (0.12) 1.01 (0.01)a 89 65.58 (0.99) 3.25 (0.17) 1.04 (0.01)b 
2004 
 May 100 40.62 (1.05) 0.83 (0.04) 0.99 (0.02)a 97 48.53 (0.95) 1.49 (0.12) 1.10 (0.02)b 99 43.15 (1.27) 1.06 (0.15) 0.99 (0.02)a 
 June 51 50.04 (1.73) 1.93 (0.23) 1.12 (0.02)a 38 55.42 (1.97) 2.11 (0.21) 1.10 (0.03)a 47 55.23 (2.11) 2.18 (0.30) 1.04 (0.02)b 
 Julya 92 53.91 (1.03) 2.00 (0.13) 1.15 (0.02) 91 58.53 (1.41) 2.79 (0.18) 1.19 (0.03) 92 51.88 (0.82) 1.70 (0.07) 1.20 (0.03) 
 Augusta 87 57.09 (0.70) 1.98 (0.08) 1.02 (0.01) 47 62.06 (1.64) 2.80 (0.23) 1.06 (0.01) 39 62.08 (1.78) 2.69 (0.26) 1.07 (0.03) 
 September 57 62.26 (1.53) 2.86 (0.29) 1.03 (0.01) 98 62.09 (0.99) 2.70 (0.06) 1.04 (0.02) 103 59.24 (0.95) 2.30 (0.15) 1.01 (0.01) 
 October 77 62.39 (1.04) 2.57 (0.17) 0.98 (0.01)a 100 66.39 (1.04) 3.14 (0.17) 0.99 (0.01)a 97 62.05 (0.98) 2.47 (0.15) 0.95 (0.01)b 

Sampling was conducted post-freshet and in ice-free months only at a constant catch per unit effort of 10,000 electrofishing seconds. All fish were measured and released back into collection areas. Data are shown as mean (standard error). Letters (a,b) indicate a significant difference in condition (length–weight relationship) between sites in that month only.

aIndicates presence of interaction in ANCOVA analysis (p ≤ 0.05).

The abundance of fish at Ten Mile (Figure 2(a)) was stable over the collection periods, ranging from 200 to 670 sculpin (per 10,000 s electroshocking), with an average (478) similar to that found at Donat (472) (Figure 2(b)). The site downstream of the agricultural area had the highest average abundance (751), and the most variability (ranging from 172 to 1972). There was much less variability at Ten Mile indicating stability of the fish populations in the areas not receiving agricultural inputs (Figure 2(a)). The abundance in the middle reaches (Figure 2(b)) is similar to the forested site, but the agricultural area is characterized by large numbers of YOY following emergence each year, with the exception of 2004. The increase is followed by a rapid decline in population size; this decline was not evident at either upstream site. Following initial emergence in 2002 and 2003, YOY densities remained high at 0.80 and 0.69 for September and October, respectively (Table 2). At the most upstream forested site, Ten Mile, August YOY densities were lower at 0.21, but upon subsequent sampling, the proportion more than doubled, suggesting the occurrence of late hatching and/or slower growth at this site.

Figure 2

Changes in slimy sculpin population structure for densities of slimy sculpin collected monthly along a gradient of agricultural intensity, (a) Ten Mile (forested), (b) Donat (intermediate), (c) Dead (Agriculture). Data shown are estimated catch based on standardized effort of 10,000 electrofishing seconds across all sites. Data are categorized into bins based on fish total length (mm).

Figure 2

Changes in slimy sculpin population structure for densities of slimy sculpin collected monthly along a gradient of agricultural intensity, (a) Ten Mile (forested), (b) Donat (intermediate), (c) Dead (Agriculture). Data shown are estimated catch based on standardized effort of 10,000 electrofishing seconds across all sites. Data are categorized into bins based on fish total length (mm).

Upstream numbers of adults were relatively stable year round, and at all sites, large sculpin (>70 mm) counts appeared relatively stable from November to June, suggesting that overwinter mortality of larger individuals was not important. At Ten Mile, YOY densities were stable or increasing during the fall in 2004, but at the downstream site, Dead, the major peak of fish numbers appeared in August, and then declined over the fall (Table 2). In 2004, Donat showed a similar pattern as Dead Brook, with an increase in YOY in August with emergence followed by a decrease in YOY in subsequent samples.

Climate data

Precipitation, reported as total annual rainfall, was lowest in 2002 at 639.2 mm with 481 mm falling during the cropping season (Table 4). As the wettest year of this study, 939.9 mm of rain fell in 2003, 38% of which occurred in July and August, with a total of 734 mm in the cropping season. In the final year of study, 2004, 706.3 mm of rain fell with 529 mm during the cropping season. Water temperature, reported as degree-days, did not vary among the three sites; however, Ten Mile had lower temperatures across most years (Table 5).

Table 4

Precipitation data reported as monthly and annual total rainfall (mm) at the St Leonard airport (Station 6256, Canada National Climate Archive) with May to October cropping season in bold

  Total precipitation
 
Storms > 15 mm/> 25 mm/> 35 mm
 
2002 2003 2004 2002 2003 2004 
January 0.2    
February 18.2 2.2    
March 22.6 31.0 23.7    
April 59.0 31.8 66.2 2/0/0 1/0/0 
May 74 81.8 70.1 0/0/0 0/0/0 0/0/0 
June 68.8 103.1 93.0 0/0/0 2/0/0 3/0/0 
July 101.2 225.8 122.3 2/0/0 3/3/3 2/2/1 
August 58.0 135.9 112.5 2/0/0 3/2/0 3/2/0 
September 116.8 37.2 80.0 2/2/0 0/0/0 2/1/0 
October 62.2 150.2 51.3 1/0/0 3/2/1 0/0/0 
November 31.8 94.2 59.8 1/0/0 4/1/1 3/0/0 
December 26.4 46.7 27.4 1/0/0 1/1/1 3/1/0 
Annual 639.2 939.9 706.3 11/2/0 16/9/6 17/6/1 
Cropping season 481 734 529 7/2/0 11/7/4 10/5/1 
  Total precipitation
 
Storms > 15 mm/> 25 mm/> 35 mm
 
2002 2003 2004 2002 2003 2004 
January 0.2    
February 18.2 2.2    
March 22.6 31.0 23.7    
April 59.0 31.8 66.2 2/0/0 1/0/0 
May 74 81.8 70.1 0/0/0 0/0/0 0/0/0 
June 68.8 103.1 93.0 0/0/0 2/0/0 3/0/0 
July 101.2 225.8 122.3 2/0/0 3/3/3 2/2/1 
August 58.0 135.9 112.5 2/0/0 3/2/0 3/2/0 
September 116.8 37.2 80.0 2/2/0 0/0/0 2/1/0 
October 62.2 150.2 51.3 1/0/0 3/2/1 0/0/0 
November 31.8 94.2 59.8 1/0/0 4/1/1 3/0/0 
December 26.4 46.7 27.4 1/0/0 1/1/1 3/1/0 
Annual 639.2 939.9 706.3 11/2/0 16/9/6 17/6/1 
Cropping season 481 734 529 7/2/0 11/7/4 10/5/1 
Table 5

Degree-days (sum of mean daily temperature above 0 °C) for water temperatures collected between 27 July and 18 October for the Little River

Degree-days above 0 °C
 
  2002 2003 2004 Mean 
Ten Mile 931.19 939.96 917.54 929.56 
Donat 997.11 1027.46 965.25 996.60 
Dead 931.09 1014.83 1024.27 990.06 
Degree-days above 0 °C
 
  2002 2003 2004 Mean 
Ten Mile 931.19 939.96 917.54 929.56 
Donat 997.11 1027.46 965.25 996.60 
Dead 931.09 1014.83 1024.27 990.06 

Temperature data for 2002 were extrapolated from a regression generated by air: water temperature from 3 other years. Air temperature data were collected at the St Leonard airport, Station ID 6256, Canada National Climate Archive.

Pesticide analyses

Surface water samples collected weekly and during rainfall-runoff events (Figure 3) at Weir 1 (Figure 1) from September 2003 to December 2004 demonstrate pesticide residues associated with potato production. Overall, the highest residues were associated with rainfall-runoff events. The highest levels detected during the course of the study occurred during two separate rain events in July 2004, where the in-stream concentrations of azinphos-methyl, imidacloprid, and linuron exceeded Canadian Council for Ministers of the Environment (CCME) water quality guidelines of 20 ng/L, 230 ng/L, and 71 ng/L, respectively (Figure 3). The highest exceedance of a CCME guideline was observed for azinphos-methyl, whose peak concentration of 9,225 ng/L exceeded its guideline by 461-fold during the July 2004 storm. Azinphos-methyl and linuron also exceeded their CCME guidelines during three other events in August and September (Figure 3), but by no more than 2.2-fold. Concentrations of the other pesticides in baseflow and event-related surface water exiting Black Brook over the course of this study were relatively minor with maximum concentrations that were well below CCME guidelines, where they exist (metobromuron maximum <1.5 ng/L, metalaxyl <145 ng/L, metribuzin <113 ng/L, carbaryl <0.1 ng/L).

Figure 3

Surface water concentrations of azinphos-methyl, imidacloprid, linuron (solid lines) and their corresponding CCME water quality guidelines (solid horizontal lines) at Weir 1 of Black Brook, from September 2003 to December 2004. Stream flow and rain events are overlaid (dotted lines).

Figure 3

Surface water concentrations of azinphos-methyl, imidacloprid, linuron (solid lines) and their corresponding CCME water quality guidelines (solid horizontal lines) at Weir 1 of Black Brook, from September 2003 to December 2004. Stream flow and rain events are overlaid (dotted lines).

Stressor correlations

In an effort to determine factors that may be associated with the changes in fish densities observed, relationships between maximum daily temperature (27 July–18 Oct) and fish abundance were considered (Figure 4); however, there was no significant relationship observed for either forested sites (p =0.49) or agricultural sites (p =0.30). Precipitation, reported as total rainfall, was summarized in late summer (July–August) and regressed against %YOY for both forested and agricultural collections. Although there was no relationship found for forested sites (p = 0.80), this relationship was significant in agricultural sites (p = 0.039) (Figure 5). To further analyze, this relationship between %YOY and precipitation, major rainfall events were defined as storms that exceeded 15 mm of total precipitation. Correlating %YOY and the number of storm events over 15 mm resulted in a non-significant relationship (Figure 6(a)). However, YOY abundance was regressed against cumulative storm events for the previous 1–3-year period (Figure 6(b)), and significant regressions were found for the total storm events over the combined previous 3 years. From the reported pesticide data, the relationships between relative fish abundance at the lowermost site (Dead) and chemical concentration were assessed. Associated with a major storm event in July 2004, the concentrations of azinphos-methyl, imidacloprid and linuron were in exceedance of their respective CCME guidelines (Figure 3), and at the Donat and Dead sites the abundance of adult sculpin following these pulses of pesticides in August were observed to be some of the lowest recorded in the course of this study (Figure 2(b) and 2(c)).

Figure 4

Relationship between maximum daily water temperature and estimated abundance of all sculpin based on standardized effort.

Figure 4

Relationship between maximum daily water temperature and estimated abundance of all sculpin based on standardized effort.

Figure 5

Linear relationship between total rainfall (mm) and percent YOY for sites along the gradient of agriculture.

Figure 5

Linear relationship between total rainfall (mm) and percent YOY for sites along the gradient of agriculture.

Figure 6

Relationships between (a) number of late summer storms > 15 mm in July and August shown as total precipitation and percent YOY sculpin (b) number of summer storms > 25 mm (from previous 1–3 years) versus YOY abundance.

Figure 6

Relationships between (a) number of late summer storms > 15 mm in July and August shown as total precipitation and percent YOY sculpin (b) number of summer storms > 25 mm (from previous 1–3 years) versus YOY abundance.

DISCUSSION

This study utilized non-lethal collection methods of slimy sculpin to demonstrate impacts associated with intensive agricultural activity in northwestern NB. Relative to upstream forested reaches, fish collected below in drainages of higher agricultural land use showed increased growth and condition, higher abundance and increased variability in abundance. Consistent with previous studies, sculpin downstream of agricultural inputs are longer and heavier than those of upstream non-agricultural sites (Gray et al. 2002; Gray & Munkittrick 2005; Gray et al. 2005). Previous studies indicated that sculpin populations within the Little River are distinct, based on stable isotopic signatures for carbon and nitrogen (Gray et al. 2004). Previous studies did not find major differences between sampling conducted on the Little River, a comparison of multiple sites along three additional rivers, or at single sites examined on more than 20 different rivers. This sampling was non-lethal and did not measure internal organ weights, but previous studies (Gray et al. 2005) have shown smaller livers and gonads, and reduced fecundity in the agriculturally influenced reaches. Our agricultural site at the convergence of the Dead Brook with the Little River appears to be relatively unimpacted relative to other reports in this system. One previous study (Gray et al. 2005) observed more than 80% of agricultural tributaries did not have successful reproduction. However, this study was conducted in the main stem of Little River and damage may have been more extensive in smaller tributaries with a higher percentage of agricultural land use.

In comparison to previous studies, this study did not find reduced densities of YOY in agricultural areas, but did record increased variability in those areas. Additionally, the previous studies recorded a decreased proportion of YOY at agricultural sites (Gray et al. 2002,  2005), while this study showed higher proportions of YOY in the population for 2 of the 3 years monitored. This study demonstrated a significant negative relationship between year class strength (proportion of the population comprised of YOY) and number of major storms (>15 mm rainfall); sculpin year classes were weaker during wetter years with more runoff events. This relationship was not seen at forested sites. Longer term studies in the southeastern USA demonstrated that over a 10-year period, abundances of most species either increased or remained unchanged during low flow, and in fact mortality from high-flow events had a stronger impact on population size than stresses imposed by low flow (Grossman et al. 1998). The YOY were larger in the sites downstream of agricultural inputs, with median sizes of 29 mm (Donat) and 30 mm (Dead) versus 25 mm in the upstream site (Ten Mile). The major consistent decrease in abundance of larger fish is between July and August (2002) or July and October (2003), suggesting that the peak mortality for both larger and smaller fish coincides with the time (June to September) when most soil erosion and runoff occur (Chow et al. 1990; Rees et al. 2002).

Our data indicate a significant relationship with the number of major rain events in July and August (precipitation exceeding 15 mm) and fish populations in streams adjacent to agricultural populations. Research related to soil loss and erosion in these systems indicates that the majority of soil losses occur following high-intensity thunderstorms during the growing season (Chow et al. 2000; Rees et al. 2002). These summer thunderstorms coincide with both chemical applications and the period of major acute fish kill events in potato-growing areas (Cairns 2002). As can be seen from the pesticide monitoring conducted during this study, the events on 23 July 2004 resulted in significant pulses of the insecticides imidacloprid (>CCME guideline) and azinphos-methyl (>CCME guideline), the herbicides metribuzin (no CCME guideline available) and linuron (>CCME guideline), and the fungicide metalaxyl (no CCME guideline available) (Figure 3, data shown for those in relation to exceeding CCME guideline only). Although the concentrations exiting Black Brook would be reduced by dilution from their transit in Dead Brook, the spikes of pesticide concentrations during rain events cannot be discounted in the relationship with weaker sculpin year classes in July and August 2004 (Figure 2).

The presence of pesticides in event-associated surface water collections is the result of many factors, including storm intensity, level of implemented soil and water conservation practices such as buffer strips, conservation tillage, diversions and grassed waterways (Chow et al. 2000), and pesticide mobility and usage patterns (Milburn et al. 1991). Maximum concentrations for nearly all pesticides measured in our study occurred during the rain events of 12 and 23 July 2004, and the hydrograph indicates that this was the first major event of the growing season (and since 3 May) following spring chemical applications; this is often referred to as the ‘first wash’ effect. Additional information on pesticide levels in the Black Brook watershed was limited. Milburn et al. (1991) measured metribuzin in tile drainage from a field near Grand Falls from 1987 to 1989 and detected a maximum concentration of 1.53 μg/L in the first event after application, and average metribuzin levels of 220 ng/L, 10 months after application. These results suggest that this particular compound was somewhat mobile, which is consistent with the data obtained in this study (Figure 3).

Temperature is higher in agricultural areas (Gray et al. 2004; Gray & Munkittrick 2005) but in this study, the median size of YOY was not different between sites along the gradient. Gray et al. (2004) reported much higher temperatures in agricultural sites in this watershed and in those cases, sculpin were found in much lower densities or entirely absent. Temperature is recognized as a major ecological factor affecting the development of freshwater species (e.g. Vannote & Sweeney 1980) and is thought to influence the abundance of fish populations through growth and fecundity (Lobón-Cerviá & Rincón 1998). The relationships between water temperature, fish growth, and recruitment success have received considerable attention (Mann et al. 1984). Temperature controls the rate of food consumption and metabolism, and thus fish growth (Nunn et al. 2003). Additionally, a recent European study of bullhead (Cottus gobio) suggests that the distribution of populations and individuals, was first structured by the suitability of physical habitat and hydraulic conditions, and then population dynamics was mainly governed by the thermal regime (Legalle et al. 2005). This would support the importance of temperature as a habitat variable in this family of fishes.

At the two upstream stations (Ten Mile and Donat), YOY densities are stable or increasing during the fall, but at the downstream site (Dead), the major peak appeared usually in August, and then showed rapid declines over the fall. Prior to this study, summer densities at these sites were reported as 400 and 500 per 10,000 s (Gray et al. 2002), these data were very similar to what was collected in 2002, but lower densities were recorded in 2004. The abundance of large sculpin was high in 2002, after consecutive dry years (Figure 6), but 2003 and 2004 had high-precipitation levels due to summer storms, and were two years with lower survival characterized by major summer storms (five in 2003 and four in 2004).

Among larger fish at the upstream sites, numbers of adult sculpin are relatively stable year round, and large sculpin (>70 mm) remained relatively constant from November to May at all sites. This suggests that overwinter mortality of larger individuals was not important as previous studies indicated. This may imply that stream baseflow associated with winter was not a factor. The major consistent decrease in abundance of larger fish is between July and August (2002) or July and October (2003), suggesting that the peak mortality for both larger and smaller fish coincided with the period of warmest water temperatures and highest rainfall events in late summer which would also have had the highest levels of soil erosion (soil loss and runoff).

Our study has confirmed previous findings of faster growth, and larger sizes of sculpin in agricultural areas, but found that year class strength and larval survival were dependent on the number and severity of summer rainfall events. This study contributes information on sculpin population changes within and between years in an impacted watershed, recent information on the reproductive timing, and population dynamics in an ice-free system in southern NB (Keeler & Cunjak 2007). Although the standardized non-lethal sampling design was improved over previous studies to isolate periods of risk to fish populations, the data remain difficult to interpret. Condition factor showed some differences in energy storage between sites, but no clear patterns emerged. Additional ecological studies were also needed to better understand the response patterns observed in slimy sculpin.

Potato-cultivating activities subject streams to a complex array of stressors associated with increased water temperatures, increased nutrients from fertilizers, increased sediment loading, increased runoff associated with storm events, and increased chemical exposures associated with pesticide – insecticide, fungicide and herbicide – applications. Based on the fact that the Black Brook baseflow pesticide levels are below their CCME guidelines, it is likely that these conditions do not present a risk to aquatic life. However, the pulses seen in July and August 2004 coincide with reduced population densities and reduced numbers of YOY. Whether these pulse-type exposures themselves are sufficient to elicit the effects observed, or whether the exposures act in combination with the rainfall, sediment loadings, and temperatures, needs further study.

Although adult sculpin experienced higher mortality during summer periods (July–August), they may not be the most sensitive fish species for evaluating reproductive impacts of agricultural impacts on the aquatic environment. Their spawning period is completed before summer storms, chemical applications and warm temperatures, and peak exposures to stress occur during a period of reproductive inactivity (Brasfield et al. 2013). In NB streams, other fish species present are relatively limited, including brook stickleback (Culea inconstans) and brook trout (Salvelinus fontinalis), and in warmer downstream areas there are also blacknose dace (Rhinichtys atratulus) (Curry & Munkittrick 2005). It would be important to evaluate these other species for potential impacts in future work.

ACKNOWLEDGEMENTS

The authors received funding from the Canadian Water Network, the New Brunswick Innovation Fund, Environment Canada's Pesticide Science Fund and Natural Sciences and Engineering Research Council (NSERC). The invaluable help of countless field assistants, graduate students, and visiting scholars is greatly appreciated. Kelly Munkittrick receives support from a NSERC Discovery Grant, the Canadian Water Network, Networks of Centres of Excellence and from the Canada Research Chairs Program.

REFERENCES

REFERENCES
Battaglin
W.
Fairchild
J.
2002
Potential toxicity of pesticides measured in Midwestern streams to aquatic organisms
.
Water Science and Technology
45
,
95
103
.
Böhlke
J. K.
2002
Groundwater recharge and agricultural contamination
.
Hydrogeological Journal
10
,
153
179
.
Bouwer
H.
1990
Agricultural chemicals and groundwater quality
.
Journal of Soil and Water Conservation
45
,
184
189
.
Brasfield
S. M.
Tetreault
G.
Bennett
J.
McMaster
M. E.
Munkittrick
K. R.
2013
Seasonal patterns of gonad size, liver size, and in vitro gonadal steroidogenic capacity in slimy sculpin (Cottus cognatus)
.
Water Quality Research Journal of Canada
48
(
3
),
243
254
.
Cairns
D. K.
2002
Effects of Land Use Practices on Fish, Shellfish, and their Habitats on Prince Edward Island
.
Canadian Technical Report of Fisheries and Aquatic Sciences No. 2408. Department of Fisheries and Oceans Canada
,
Ottawa, ON, Canada
.
Chow
T. L.
Rees
H. W.
1995
Effects of coarse-fragment content and size on soil erosion under simulated rainfall
.
Canadian Journal of Soil Science
75
,
227
232
.
Chow
T. L.
Daigle
J.-L.
Ghanem
I.
Cormier
H.
1990
Effects of potato cropping practices on water runoff and soil erosion
.
Canadian Journal of Soil Science
70
,
137
148
.
Clark
G. M.
Goolsby
D. A.
Battaglin
W. A.
1999
Seasonal and annual load of herbicides from the Mississippi River basin to the Gulf of Mexico
.
Environmental Science and Technology
33
,
981
986
.
Culliney
T. W.
Pimentel
D.
Pimentel
M. H.
1992
Pesticides and natural toxicants in food
.
Agriculture, Ecosystems and the Environment
41
,
297
320
.
Cunjak
R. A.
Roussel
J.-M.
Gray
M. A.
Dietrich
J. P.
Cartwright
D. F.
Munkittrick
K. R.
Jardine
T. D.
2005
Using stable isotope analysis with telemetry or mark-recapture data to identify fish movement and foraging
.
Oecologia
144
,
636
646
.
Curry
R. A.
Munkittrick
K. R.
2005
Fish community responses to multiple stressors along the Saint John River, New Brunswick, Canada
. In
Changes in Large River Fish Assemblages in the North America: Implications for Management and Sustainability of Native Species
, (
Rinne
J. N.
Calamusso
R.
Hughes
R.
, eds).
American Fisheries Society
,
Bethesda, MD, USA
, pp.
505
521
.
Gray
M. A.
Munkittrick
K. R.
2005
An effects-based assessment of slimy sculpin (Cottus cognatus) populations in agricultural regions of northwestern New Brunswick
.
Water Quality Research Journal of Canada
40
,
16
27
.
Gray
M. A.
Curry
R. A.
Munkittrick
K. R.
2002
Non-lethal sampling methods for assessing environmental impacts using a small-bodied sentinel fish species
.
Water Quality Research Journal of Canada
37
,
195
211
.
Gray
M. A.
Cunjak
R. A.
Munkittrick
K. R.
2004
Site fidelity of slimy sculpin (Cottus cognatus): Insights from stable carbon and nitrogen analysis
.
Canadian Journal of Fisheries and Aquatic Sciences
61
,
1717
1722
.
Gray
M. A.
Curry
R. A.
Munkittrick
K. R.
2005
Impacts of nonpoint inputs from potato farming on populations of slimy sculpin (Cottus cognatus)
.
Environmental Toxicology and Chemistry
24
,
2291
2298
.
Grossman
G. D.
Ratajczak
R. E.
Crawford
M.
Freeman
M. C.
1998
Assemblage organization in stream fishes: effects of environmental variation and interspecific interactions
.
Ecological Monographs
68
,
395
420
.
Hennion
M. C.
2000
Graphitized carbons for solid-phase extraction
.
Journal of Chromatography A
885
,
73
95
.
Karr
J. R.
1999
Defining and measuring river health
.
Freshwater Biology
41
,
221
234
.
Keeler
R. A.
Cunjak
R. A.
2007
Reproductive ecology of slimy sculpin in small New Brunswick streams
.
Transactions of the American Fisheries Society
136
,
1762
1768
.
Legalle
M.
Santoul
F.
Figuerola
J.
Mastrorillo
S.
Ce Re Ghino
R.
2005
Factors influencing the spatial distribution patterns of the bullhead (Cottus gobio L., Teleostei Cottidae): a multi-scale study
.
Biodiversity and Conservation
14
,
1319
1334
.
Mann
R. H. K.
Mills
C. A.
Crisp
D. T.
1984
Geographical variation in the life-history of some species of freshwater fish
. In:
Fish Reproduction: Strategies and Tactics
(
Potts
G. W.
Wootton
R. J.
, eds).
Academic Press
,
London
, pp.
171
186
.
Milburn
P.
O'Neill
H.
Fartlye
C.
Pollock
T.
Richards
J. E.
Bailey
H.
1991
Leaching of dinoseb and metribuzin from potato fields in New Brunswick
.
Canadian Agricultural Engineering
33
,
197
204
.
Munkittrick
K. R.
McMaster
M. E.
Van Der Kraak
G.
Portt
C.
Gibbons
W.
Farwell
A.
Gray
M. A.
2000
Development of Methods for Effects-Based Cumulative Effects Assessment Using Fish Populations: Moose River Project
.
SETAC Press
,
Pensacola, FL, USA
.
Napier
T. L.
Brown
D. E.
1993
Factors affecting attitudes toward groundwater pollution among Ohio farmers
.
Journal of Soil and Water Conservation
48
,
439
439
.
Pimental
D.
Harvey
C.
Resosudarmo
P.
Sinclair
K.
Kurz
D.
McNair
M.
Crist
S.
Shpritz
L.
Fitton
L.
Saffouri
R.
Blair
R.
1995
Environmental and economic costs of soil erosion and conservation benefits
.
Science
267
,
1117
1123
.
Potter
K. M.
Cubbage
F. W.
Blank
G. B.
Schaberg
R. H.
2004
A watershed-scale model for predicting nonpoint pollution risk in North Carolina
.
Environmental Management
34
,
62
74
.
Rees
H. W.
Chow
T. L.
Loro
P. J.
Lavoie
J.
Monteith
J. O.
Blaauw
A.
2002
Hay mulching to reduce runoff and soil loss under intensive potato production in northwestern New Brunswick, Canada
.
Canadian Journal of Soil Science
82
,
249
258
.
Waters
T. F.
1995
Sediment in streams: sources, biological effects, and control
.
American Fisheries Society Monograph
7
.
Welch
H. E.
Symons
P. E. K.
Narver
D. W.
1977
Some Effects of Potato Farming and Forest Clearcutting on Small New Brunswick Streams
.
Technical Report 745. Fisheries and Marine Service, St
.
Andrews, NB, Canada
.
Werner
E. E.
Gilliam
J. F.
1984
The ontogenetic niche and species interactions in size-structured populations
.
Annual Review of Ecology Systematics
15
,
393
425
.
Wischmeier
W. H.
Smith
D. D.
1978
Predicting Rainfall Erosion Losses – A Guide to Conservation Planning
.
US Department of Agriculture
,
Agriculture Handbook No. 537
.