Mercury Trends in Predatory Fish in Great Slave ... - ACS Publications

Oct 10, 2013 - Here we investigate long-term trends in Hg concentrations in Great Slave Lake (GSL) located in the Northwest Territories (NWT) of Canad...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Mercury Trends in Predatory Fish in Great Slave Lake: The Influence of Temperature and Other Climate Drivers Marlene Evans,*,† Derek Muir,‡ Robert B. Brua,§ Jonathan Keating,† and Xiaowa Wang‡ †

Aquatic Contaminants Research Division, Environment Canada, Saskatoon, Saskatchewan, Canada S7N 3H5 Aquatic Contaminants Research Division, Environment Canada, Burlington, Ontario, Canada L7R 4A6 § Watershed Hydrology and Ecology Research Division, Environment Canada, Saskatoon, Saskatchewan, Canada S7N 3H5 ‡

S Supporting Information *

ABSTRACT: Here we report on trends in mercury (Hg) concentrations in lake trout (Salvelinus namaycush), burbot (Lota lota), and northern pike (Esox lucius) from Great Slave Lake, located in the Mackenzie River Basin (MRB) and investigate how climate factors may be influencing these trends. Hg concentrations in lake trout and burbot increased significantly over the early 1990s to 2012 in the two major regions of the lake; no trend was evident for northern pike over 1999−2012. Temporal variations in Hg concentrations in lake trout and burbot were similar with respect to timing of peaks and troughs. Inclusion of climate variables based on annual means, particularly temperature, improved explanatory power for variations in Hg over analyses based only on year and fish length; unexpectedly, the temperature coefficient was negative. Climate analyses based on growing season means (defined as May−September) had less explanatory power suggesting that trends were more strongly associated with colder months within the year. Inclusion of the Pacific/North American index improved explanatory power for the lake trout model suggesting that trends may have been affected by air circulation patterns. Overall, while our study confirmed previously reported trends of Hg increase in burbot in the MRB, we found no evidence that these trends were directly driven by increasing temperatures and productivity.



INTRODUCTION Human activity has increased the amount of mercury (Hg) circulating in the environment; evidence for this is preserved in sediment and other archives.1−3 Future trends with respect to Hg concentrations in northern biota are uncertain. While North America and Europe have reduced their Hg emissions, these reductions may be counterbalanced by the rapidly accelerating growth of Asian sources which are now a major source of Hg to North America.4−7 Moreover, global warming is occurring at a particularly rapid rate in northern Canada,8,9 potentially resulting in numerous changes in aquatic ecosystem functioning including productivity and contaminant pathways.8,10−13 Warming can also affect other aspects of local climate, potentially affecting Hg trends. Wind speeds are decreasing while precipitation trends are more variable.14,15 Precipitation affects wet deposition rates and hence the delivery of Hg to watershed and lake surfaces.16 Wind speed has many effects including: Hg dry deposition and evasion rates; thermal stratification and water mixing; and Hg methylation rates in the particulate-rich thermocline of the stratified water column and the sediment−water interface.17 Fluctuations in major air circulation patterns have profound effects on local climate.18−21 Positive Pacific/North American (PNA) indices are associated with warm temperatures and an Published 2013 by the American Chemical Society

enhanced high pressure ridge over western Canada which interrupts the southern flow of cold Arctic air;22 negative Arctic oscillation (AO) indices are associated with high pressure in the polar regions and a strong flow of cold Arctic air to lower latitudes.23 While atmospheric Hg is largely in the elemental form (Hg°), it is oxidized in the atmosphere with reactive gaseous Hg concentrations increasing with altitude;24−26 subsidence of these air masses can be a major source of Hg2+ at sites distant from sources5 where it can become incorporated into aquatic and marine ecosystems through various pathways.24 Such processes may be particularly important when air is forced to rise, for example, the orthogonal rise of Asian air masses over the mountains of western Canada. While Hg concentrations have been measured in biota from many Arctic and sub-Arctic locations over time, the frequency of monitoring in most studies has been inadequate to detect trends.27 One exception is the burbot (Lota lota) monitoring at Fort Good Hope (FGH) on the Mackenzie River (MR), Canada which has detected an increasing trend of Hg.10 The Received: Revised: Accepted: Published: 12793

June 14, 2013 October 1, 2013 October 10, 2013 October 10, 2013 dx.doi.org/10.1021/es402645x | Environ. Sci. Technol. 2013, 47, 12793−12801

Environmental Science & Technology

Article

Figure 1. Upper panel; map of Great Slave Lake showing major communities and locations of sediment core sites. Lower panels: Hg wet weight concentrations (mean and ± standard error) in burbot and lake trout collected from the West Basin and East Arm of Great Slave Lake; northern pike were collected near Fort Resolution.

between temperature and inferred productivity trends. We compared our findings with those observed for Hg trends in FGH burbot.10

specific mechanism proposed driving this increase was a temperature-driven enhancement in productivity and Hg scavenging. However, increased productivity in the water column can reduce bioavailability of Hg to pelagic fish and remove Hg from the water column with increased sedimentation.28,29 Here we investigate long-term trends in Hg concentrations in Great Slave Lake (GSL) located in the Northwest Territories (NWT) of Canada. Hg concentrations have been regularly monitored in lake trout (Salvelinus namaycush) in GSL at two contrasting locations since 1999 and northern pike (Esox lucius) periodically at one location; earlier data are also available for fish Hg concentrations.30,31 These data sets allowed us to investigate time trends in Hg concentrations in three predatory species with different ecological niches inhabiting two limnologically distinct regions of GSL. We first considered fish length and year as influencing variables and then the role of climate variability considering not only temperature but precipitation and wind speed. PNA and AO indices were considered given their importance in affecting regional climate and the potential for PNA to affect atmospheric Hg oxidation rates at higher altitudes over western Canada. Finally, sediment cores were used to investigate the long-term record of sediment Hg concentrations and flux to GSL, as well as, the relationship



MATERIALS AND METHODS Study Areas. GSL (Figure 1), the world’s ninth largest lake (with a surface area of 28 568 km2 and a drainage area of 983 000 km2) is located in the Mackenzie River Basin (MRB) and forms the headwaters of the Mackenzie River;32 to the west are the Cassiar Mountains, northern extensions of the Rocky Mountain chain. GSL is divided into the relatively productive West Basin (WB) which has supported a commercial fishery since the late 1940s and the oligotrophic East Arm (EA); the WB is profoundly influenced by the Slave River (SR) which contributes about 87% to the water budget of the lake (31 000−52 000 m3/yr) and 2.64 to 6.72 × 1010 kg/yr of suspended sediment.33 Further details on the lake and the species monitored are given in the Supporting Information (SI). Fish Collections. The formal Hg monitoring program began in 1999 with the three species collected in two regions of the lake; over the years, budget cuts forced a reduction in the extent of the trend monitoring with some components resumed when alternate funding was located. Lake trout were collected annually from the EA (Lutsel K’e (LK)) and the WB (Hay 12794

dx.doi.org/10.1021/es402645x | Environ. Sci. Technol. 2013, 47, 12793−12801

Environmental Science & Technology

Article

Table 1. Biological Attributes and Average Mercury Concentrations in Muscle of Lake Trout, Burbot, and Northern Pike from the West Basin (Fort Resolution) and East Arm (Lutsel K’e)a location Lake Trout West Basin East Arm: all East Arm: 640 mm Burbot West Basin East Arm Northern Pike West Basin

total length (mm) 648.3 641.8 590.9 697.6

± ± ± ±

66.5 66.4 36.5 42.3

weight (g) 2891 2489 1916 3199

± ± ± ±

998 851 352 790

age (yr) 8.3 14.0 13.4 15.0

± ± ± ±

3.0 2.8 2.5 2.8

δ13C (‰) −29.6 −28.4 −27.8 −28.8

± ± ± ±

1.0 1.8 1.8 1.3

δ15N (‰) 12.5 11.7 11.7 11.9

± ± ± ±

0.6 0.6 0.6 0.5

Hg (μg/g wet wt.) 0.18 0.15 0.13 0.18

± ± ± ±

0.08 0.09 0.06 0.08

657.4 ± 65.7 533.8 ± 61.2

2006 ± 606 1119 ± 419

12.9 ± 3.0 10.4 ± 2.4

−28.5 ± 1.2 −23.7 ± 2.4

12.0 ± 0.9 11.2 ± 1.0

0.15 ± 0.06 0.13 ± 0.05

682.4 ± 73.4

2,705 ± 942

9.0 ± 2.3

−28.4 ± 0.7

11.0 ± 0.8

0.24 ± 0.10

a

Lake trout from the West Basin (Hay River) and East Arm (Lutsel K’e) over 1993−2012 and northern pike from the West Basin (Fort Resolution) over 1992−2012. Data are shown as simple mean ± one standard deviation.

Cores, collected with a 7.6 cm diameter benthic corer, were sliced at 0.5 cm intervals down to 10 cm, 1 cm intervals down to 20 cm and 2 cm intervals thereafter. Cores were dated and sedimentation rates were determined using 210Pb and the constant rate of supply model.38 Methods for storage and dating of the cores are provided in the SI. Total Hg analyses were performed on a subset of freeze-dried core slices by CVAAS. Hg flux was calculated as the product of sedimentation rate and Hg concentration. Organic carbon was determined by using a CHN analyzer (see Muir et al.2) and organic carbon fraction S2, a measure of algal carbon, and RC, a measure of refractile carbon, were determined by Rock-Eval analyses.39 Climate and Arctic Oscillation and Pacific/North American Indices. Climate variability and trends were examined for two stations on the Great Slave Lake shoreline, HR on the south, Yellowknife (YK) on the north shore and at Norman Wells (NW) near the site of burbot monitoring at FGH.10 Data were obtained from Environment Canada’s Climate Change web site for adjusted and homogenized data (http://www.ec.gc.ca/dccha-ahccd/default.asp? lang=En&n=B1F8423A-1). Temperature and precipitation data for NW and YK were available up to 2012, whereas data for NW were incomplete for 2012. Long-term trends were investigated over 1975−2012 and short-term over 1990− 2012; long-term climate trends provided for an assessment of climate influences over the longest GSL lake trout (1978− 2012) and MR burbot (1985−2008); the short-term period focused on the more intensive period of Hg monitoring on GSL and the MR. Data were examined as annual and MaySeptember (M-S) means with the latter period corresponding to the period in which annual mean air temperature was >0 °C and in which most aquatic productivity was assumed to occur. AO and PNA data (1950−2012) were obtained from NOAA (http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/teleconnections.shtml) and examined as annual and M-S means. Statistical Analyses. General linear model (GLM) analysis of variance was used to investigate the factors affecting variability in Hg concentrations in fish; Hg data were log10transformed to meet the assumptions of normality. For the first series of runs, only fish length, year and a year·length interaction term were considered; only fish 500−800 mm in total length were used. Other variables such as fish age and carbon and nitrogen isotope ratios were not used because they had not been measured in earlier collections and the focus of this paper is on climate factors. Analyses were run for 1999−

River, (HR)) from 1999−2002 and 2004−2012. Burbot were collected annually from the WB in Resolution Bay (RB) over the same period while in the EA, they were collected from 1999−2002 and from 2008−2012. Northern pike were collected from RB over 1999−2002 and from 2008−2012. Lake trout and pike were collected with 4.5 in. mesh gill nets while burbot generally were collected by hook and line; twenty fish of each species, with a requested size range of 500−800 mm total length were provided annually by local community or commercial fishers. This range was selected to reduce the influence of fish length on the analyses; other monitoring studies constrain fish length.34 Limited data from our burbot, lake trout and pike studies were available for 1993, 1995, and 1996 and the WB commercial fishery.30 Periodic lake trout records exists for 1979−1990. While commercial fish records exist for pike for this period, they were not available for the RB area and were not used given the sedentary nature of northern pike. Burbot data from the commercial fishery were available for 1992−1996; since this species migrates to the RB area in the fall from other regions of the WB, these data were also included. Finally, burbot and pike data were available for collections made in 1992 and 1993 in RB by Lafontaine et al.35 See Supporting Information for additional details. Laboratory Methods. Fish were shipped whole and frozen to Environment Canada (Saskatoon) where they were stored at −40 °C until processing. Total and fork length (total length only for burbot), weight and sex were determined; fish were aged using their otoliths or cleithra (pike). Stable carbon (δ13C) and nitrogen (δ15N) isotope ratios were determined on dried, nonlipid extracted homogenized fish muscle. Subsets of 10 fish, typically ranging in total length from 500 to 800 mm and, with equal numbers of males and females where possible, were selected for Hg analyses. Total Hg analyses on wet fillet were performed by tissue digestion and cold vapor atomic absorption spectrophotometry (CVAAS) until 2004 and thereafter by automated thermal decomposition and atomic absorption detection using EPA method 7473.36 Further details on analytical methods and quality assurance are provided in the SI. Sediment Cores. Sediment cores were collected in March 2009 at two sites in the WB which previously had been shown to provide undisturbed sediment records.37 Site 12, offshore of HR, was in 69 m of water in the general vicinity of where lake trout were commercially harvested; site 19, to the west and in 47 m of water, was under a stronger SR influence (Figure 1). In the EA, a core was collected in 44 m of water in Portage Inlet. 12795

dx.doi.org/10.1021/es402645x | Environ. Sci. Technol. 2013, 47, 12793−12801

Environmental Science & Technology

Article

Table 2. Results of General Linear Model Analyses of Factors Affecting Variability in Log Hg Concentration in Burbot, Lake Trout and Northern Pike (West Basin only) from the West Basin and East Arm of Great Slave Lake over Different Time Intervalsa species Lake Trout all fish all fish all fish 590 mm >590 mm Burbot all fish all fish all fish Northern Pike

time period West Basin 1999−2012 1990−2012 1979−2012 East Arm 1999−2012 1999−2012 1993−2012 West Basin 1999−2012 1992−2012 East Arm 1999−2012 West Basin 1999−2012 1992−2012

N

equation

F-ratio

R2

p

107 123 179

−18.0481 + 0.0081·Yr + 0.0016·L −18.2652 + 0.0082·Yr + 0.0018·L −13.9063 + 0.0060·Yr + 0.0017·L

69.29 92.42 132.5

0.57 0.60 0.60