Effect of eutrophication on mercury (Hg) dynamics in subtropical reservoirs from a high Hg deposition ecoregion
2015; Wiley; Volume: 60; Issue: 2 Linguagem: Inglês
10.1002/lno.10036
ISSN1939-5604
AutoresN. Roxanna Razavi, Mingzhi Qu, Dongmei Chen, Yang Zhong, Wenwei Ren, Yuxiang Wang, Linda M. Campbell,
Tópico(s)Heavy metals in environment
ResumoLimnology and OceanographyVolume 60, Issue 2 p. 386-401 ArticleFree Access Effect of eutrophication on mercury (Hg) dynamics in subtropical reservoirs from a high Hg deposition ecoregion N. Roxanna Razavi, Corresponding Author N. Roxanna Razavi Department of Biology, Queen's University, Kingston, Ontario, CanadaCorrespondence: [email protected]Search for more papers by this authorMingzhi Qu, Mingzhi Qu Department of Biology, Queen's University, Kingston, Ontario, CanadaSearch for more papers by this authorDongmei Chen, Dongmei Chen Department of Biology, Queen's University, Kingston, Ontario, CanadaSearch for more papers by this authorYang Zhong, Yang Zhong Department of Biology, Tibet University, Lhasa, China Institute of Biodiversity Science, School of Life Sciences, Fudan University, Shanghai, ChinaSearch for more papers by this authorWenwei Ren, Wenwei Ren Key Laboratory of Yangtze River Water Environment, Ministry of Education, College of Environmental Science and Engineering, Tongji University, Shanghai, ChinaSearch for more papers by this authorYuxiang Wang, Yuxiang Wang Department of Biology, Queen's University, Kingston, Ontario, CanadaSearch for more papers by this authorLinda M. Campbell, Linda M. Campbell Department of Environmental Science, Saint Mary's University, Halifax, Nova Scotia, CanadaSearch for more papers by this author N. Roxanna Razavi, Corresponding Author N. Roxanna Razavi Department of Biology, Queen's University, Kingston, Ontario, CanadaCorrespondence: [email protected]Search for more papers by this authorMingzhi Qu, Mingzhi Qu Department of Biology, Queen's University, Kingston, Ontario, CanadaSearch for more papers by this authorDongmei Chen, Dongmei Chen Department of Biology, Queen's University, Kingston, Ontario, CanadaSearch for more papers by this authorYang Zhong, Yang Zhong Department of Biology, Tibet University, Lhasa, China Institute of Biodiversity Science, School of Life Sciences, Fudan University, Shanghai, ChinaSearch for more papers by this authorWenwei Ren, Wenwei Ren Key Laboratory of Yangtze River Water Environment, Ministry of Education, College of Environmental Science and Engineering, Tongji University, Shanghai, ChinaSearch for more papers by this authorYuxiang Wang, Yuxiang Wang Department of Biology, Queen's University, Kingston, Ontario, CanadaSearch for more papers by this authorLinda M. Campbell, Linda M. Campbell Department of Environmental Science, Saint Mary's University, Halifax, Nova Scotia, CanadaSearch for more papers by this author First published: 30 January 2015 https://doi.org/10.1002/lno.10036Citations: 25AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Abstract Eutrophication can have opposite effects on mercury (Hg) bioavailability in aquatic systems, by increasing methylmercury (MeHg) production but reducing Hg biomagnification. Globally, the effect of eutrophication on Hg dynamics remains largely untested at lower latitudes such as eastern China, a productive subtropical ecoregion with Hg emission and deposition rates that are among the highest worldwide. Here, we quantify Hg (both MeHg and total Hg, THg) concentrations, Hg bioaccumulation, and Hg biomagnification rates in reservoir food webs across a gradient of eutrophication indicated by chlorophyll a (Chl a), zooplankton density, and total phosphorus (TP). We also assess the effect of hydrogeomorphic (HGM) features on Hg concentrations in water and biota. Water THg concentrations were strongly correlated with TP and were greater in reservoirs at higher elevations with short water retention times (WRT). Zooplankton and top predator THg concentrations were negatively correlated with Chl a, suggesting algal biodilution; evidence for zooplankton density dilution was also found when subtropical reservoirs were compared at a global scale with temperate lakes. Mercury bioaccumulation and biomagnification factors, respectively, did not correlate with increasing Chl a or zooplankton density suggesting no effect of plankton density on Hg trophic transfer. In subtropical reservoirs, eutrophication is associated with lower Hg concentrations in biota but does not explain Hg biomagnification; HGM features (i.e., elevation, WRT) and land use (i.e., % crop) appear to also influence Hg concentrations and bioaccumulation in reservoir food webs. Mercury (Hg) is a pollutant that threatens global fisheries due to the toxicity of its organic form, methylmercury (MeHg). Aquatic ecosystems are sites of MeHg production due to geochemical conditions that increase biological methylation by sulphate-reducing bacteria (Lin et al. 2012). From tropical to Arctic latitudes, the process of MeHg transfer across trophic levels depends on food web structure (Lavoie et al. 2013). Through changes in water geochemistry and increases in primary productivity, cultural (i.e., anthropogenic) eutrophication alters food web structure and has the potential to significantly change MeHg bioavailability and biomagnification. Conditions characteristic of eutrophication, such as low hypolimnetic dissolved oxygen concentrations and high dissolved nutrients and sediment organic matter content, promote in situ production of MeHg (Gray and Hines 2009). However, the visible hallmark of eutrophication, increased algal growth, leads to decreases in MeHg biomagnification through phytoplankton or zooplankton density dilution (Chen and Folt 2005). There remain important questions about what determines the outcome of eutrophication and Hg dynamics. It is also not clear how the changes in MeHg bioavailability and increases in plankton densities may affect the process and rate of Hg transfer between trophic levels and through the food web. Furthermore, most field studies that have measured the effect of eutrophication on Hg dynamics have focused on temperate lakes. This represents a global gap in knowledge regarding the fate of Hg with eutrophication at lower, more productive latitudes such as the subtropics. Eutrophication is often assessed either by the concentrations of total phosphorus (TP) or chlorophyll a (Chl a) in the water column. Using these indicators of eutrophication to categorize waterbodies as eutrophic or oligotrophic, authors have documented confounding patterns in MeHg production and Hg bioaccumulation with eutrophication. In temperate and subtropical freshwater systems, rates of Hg methylation were higher in more eutrophic reservoirs, increasing the flux of MeHg to the water column (Gray and Hines 2009; Meng et al. 2010). Despite the possibility for enhanced MeHg availability in water, in natural and experimentally enriched systems Hg concentrations in fish were significantly lower than in oligotrophic lakes (Essington and Houser 2003; Chen and Folt 2005). These lower Hg concentrations in upper trophic level organisms are most often attributed to algal biodilution, the partitioning of Hg into a greater amount of biomass, as suggested by negative correlations between Hg concentrations in biota and Chl a (Pickhardt et al. 2002; Chen and Folt 2005). Higher zooplankton densities can also decrease Hg transfer to fish through zooplankton density dilution (Chen and Folt 2005). Other possible mechanisms that can lower fish Hg concentrations in eutrophic systems include shifts in algal cell size (Pickhardt and Fisher 2007) and growth dilution in zooplankton and fish (Simoneau et al. 2005; Karimi et al. 2007). However, lower fish Hg concentrations with eutrophication are not always observed as evidenced by higher Hg concentrations in top predators from temperate eutrophic reservoirs (Stone et al. 2011). Using indicators of eutrophication to predict how nutrient enrichment will affect Hg dynamics is a critical research need to identify fisheries that are at risk of Hg contamination. It is important to assess not only the effect of eutrophication on Hg concentrations but also the effects on Hg biomagnification between trophic levels (i.e., predator–prey pairs) and across the food web because these processes ultimately determine Hg transfer and can be used for bioaccumulation assessments (Jardine et al. 2013). Whether or not increases or decreases in Hg concentrations, Hg bioaccumulation, and Hg biomagnification rates are expected under scenarios of greater eutrophication remains unclear. This question is especially important in reservoirs that differ markedly from lakes, for example, through water level fluctuations that can increase MeHg bioavailability (Evers et al. 2007) and external nutrient loading. Lower latitudes also deserve attention due to differences in productivity and food web structure compared to temperate regions. For example, in reservoirs of eastern China, intensive stocking of planktivorous Bighead Carp (Hypophthalmichthys nobilis) and Silver Carp (Hypophthalmichthys molitrix) can have substantial effects on plankton density and species composition, and indirectly affect trophic status of a waterbody (Li and Xu 1995). In this study, we take a broad-scale, multireservoir approach to examine the link between Hg dynamics and eutrophication in subtropical reservoirs of eastern China. China has among the highest Hg emission and atmospheric deposition rates in the world. It is estimated that deposition rates in China are 1–2 times and 1–2 magnitudes higher in rural and urban areas, respectively, compared to North America and Europe (Fu et al. 2012). Thus, the present study is akin to a natural "high exposure" treatment relative to studies in other parts of the world where this question has been previously addressed. We assess if measures of eutrophication suggested by Chen and Folt (2005), namely phytoplankton and zooplankton density, could be used to predict Hg concentrations in subtropical reservoirs. We test the predictive ability of indicators of eutrophication (Chl a, zooplankton density, and TP) to explain Hg (total Hg [THg] and MeHg) concentrations in the food web, represented by zooplankton, Bighead Carp, and top predatory fishes. Similarly, we ask whether those indicators predict derived variables describing Hg bioaccumulation (i.e., bioaccumulation factor, BAF), Hg biomagnification between a predator and prey pair (i.e., biomagnification factor, BMF), or Hg biomagnification rates (i.e., slope of the log Hg vs. δ15N regression) in food webs composed of stocked and wild fishes with various feeding ecologies. Finally, we examine how landscape features, such as hydrogeomorphic (HGM) characteristics at the basin and catchment scale, and land use, predict Hg concentrations in water, zooplankton, and top predators in these subtropical reservoir ecosystems. Our hypotheses regarding the direction of these effects are presented in Table 1. Overall, we find evidence for low MeHg availability in these subtropical reservoirs of eastern China despite their exposure to factors that are known to contribute to elevated MeHg such as elevated inputs of atmospheric Hg deposition, continued water level fluctuations, and susceptibility to eutrophication, all of which can lead to higher MeHg production. This suggests that these factors may be outweighed by biomass or growth dilution associated with eutrophication in subtropical reservoirs. Table 1. Expected outcomes of mercury (Hg) concentrations in water and biota (zooplankton and fish), Hg bioaccumulation factors (BAFbiota), and Hg biomagnification (biomagnification factor, BMF or trophic magnification factor, TMF) against effect of eutrophication (i.e., chlorophyll a (Chl a), total phosphorus (TP), or zooplankton density) and/or hydrogeomorphic (HGM) features at the catchment (HGM:catchment) and basin (HGM:basin) scale. Response variable Predictor category Hypothesized mechanism Predictor variable(s) and hypothesized direction of effect References Hgbiota Trophic status Plankton density and somatic growth dilution Chl a, zooplankton density, TP − 1,2 Higher MeHg production with increasing carbon availability and bacterial activity + 3 BAFbiota Lower MeHg assimilation efficiency with increasing carbon, higher DOC − 4,5 Higher MeHg availability due to higher MeHg production + 6 BMF, TMF Mechanisms proposed above resulting in higher or lower Hgbiota ± 7 Hgwater Land use Higher % forest cover (lower % crop cover) scavenges atmospheric Hg % Crop − 8 HGM:catchment Higher atmospheric deposition with larger catchment to surface area Catchment:surface area (CA : SA) + 9 Higher water temperatures promote methylation reservoir volume − 10 HGM:basin Higher nutrient and contaminant input Water retention time (WRT) − 11 Higher atmospheric deposition Elevation + 12 Hgbiota Land use Higher nutrient input % Crop ± 13 HGM : catchment Higher atmospheric deposition Catchment:surface area (CA : SA) − 9 Higher fish metabolic rates in warmer water Reservoir volume − 10 HGM : basin Higher nutrient and contaminant input Water retention time (WRT) − 11 Higher atmospheric deposition Elevation + 12 1Chen and Folt 2005; 2Essington and Houser 2003; 3Lin et al. 2012; 4Tsui and Wang 2004; 5Gorski et al. 2008; 6Stewart et al. 2008; 7Lavoie et al. 2013; 8Evers et al. 2007; 9Drevnick et al. 2012; 10Bodaly et al. 1993; 11Kalff 2002; 12Dittman and Driscoll 2009; 13Bremigan et al. 2008. Methods Reservoir descriptions All seven reservoirs in this study (Fig. 1) were located within the same humid to semihumid low-altitude (< 1000 m) ecoregion in eastern China (Zhou and Zheng 2008). This ecoregion has very high population densities and rapid development that have increased nutrient transport to aquatic ecosystems and resulted in the eutrophication of 85% of monitored lakes by 1996 (Zhou and Zheng 2008). The difficulties of managing fisheries in the many hypereutrophic lakes in the region have shifted the focus of fish production to reservoirs, largely because of their oligotrophic to mesotrophic statuses. To ensure that sampling did not occur during the initial period after impoundment when fish Hg concentrations are known to be most elevated, all selected reservoirs for this study were over 30 years old at the time of sampling (construction completed in 1954–1969). Reservoirs were sampled in the Summer (between late June 2011 and August 2011) and ranged in morphology from deep river—valley (e.g., F, M; Fig.1; Table 2) to shallow lake-plain types (e.g., S, T; Li and Xu 1995) and spanned a wide gradient of surface areas and catchment area to surface area ratios (CA : SA; Table 2). We strove to sample as wide a gradient of eutrophication as possible; however, access to eutrophic reservoirs was limited. Additional information regarding annual water level fluctuations, and water stratification patterns, were also not readily available. We are only aware of a detailed study on stratification patterns for one reservoir sampled here (i.e., Q), which was described as warm monomictic, with a period of thermal stratification found from April until January (Zhang et al. 2014). Two other reservoirs also stratify (i.e., Shu 1964; Zhang et al. 2004) but no data were available for other reservoirs. To the best of our knowledge, there are no direct point sources of Hg into the reservoirs; while there are coal-fired power plants in the direct vicinity (∼ 50 km2 radius) of most reservoirs (except Q and M), within a larger radius (∼ 100 km2) there are between 10 and 15 (i.e., F, M, L, Q) and ∼ 30 (i.e., H, T, S) coal-fired power plants of varying capacities, representing a substantial nonpoint source of Hg. Figure 1Open in figure viewerPowerPoint Location of reservoirs sampled in eastern China. (A) Dark gray area indicates watershed area and (B) colors indicate land use. Reservoir codes are given in parenthesis. Note: reservoirs are not to scale in the lower panel. Table 2. Physical, chemical, and biological characteristics of reservoirs sampled in eastern China (see Fig. 1 for reservoir names and locations). Population density (per km2), surface area (SA, km2), catchment area to surface area ratio (CA: SA), reservoir volume (× 109 m3), mean water level (i.e., elevation, m), and water retention time (WRT, days) were taken from published literature. Total phosphorus (TP, mg L−1), total nitrogen (TN, mg L−1), chemical oxygen demand (COD, mg L−1), Secchi depth (SeD, m), chlorophyll a (Chl a, μg L−1), zooplankton (zoop) density (ind. L−1) and biomass (μg L−1), and % forest and crop cover measured in this study. Reservoir code Population densitya SA CA: SA Volume Elevation WRTb TP TN COD SeD Chl ac Zoop density Zoop biomass % Forest % Crop F 15 20 92.0 0.364 124 100 0.069d 1.09e 2.4f 5.0 1.9 6.7 142.5 85.4 10.9 H 158 4.5 34.4 0.112 32 460 0.017 2.45 2.7 4.0 2.9 14.5 62.2 43.9 41.6 L 45 50 22.2 0.516 68 335 0.078 1.13 3.7 2.0 8.2 8.1 58.5 70.7 24.8 M 78 62.9 31.3 1.245 126 323 0.087 1.21 2.8 3.5 1.8 3.8 79.2 77.4 16.2 Q 149 580 18.1 17.84 37 700 0.018 0.62 1.3 5.5 1.5 8.0 71.4 64.2 24.4 S 324 5.5 18.7 0.0795 16 484 0.012 0.68 3.2 1.4 6.5 8.9 66.7 62.6 19.5 T 423 6.7 22.2 0.109 19 580 0.020 1.14 3.6 1.1 23.7 4.0 30.9 35.5 46.8 a Estimate based on total population in the catchment divided by the catchment area. b WRT = reservoir volume: mean outflow rate. c All samples represent means of replicate samples. d Mean of two samples TP range = 0.067–0.71. e Mean of two samples TN range = 1.04–1.13. f Mean of two samples COD range = 2.1–2.6. The percent of the catchment covered by forests vs. crops was determined using Geographic Information System (ArcGIS Version 10.1). Base maps of watersheds were obtained from BaseCamp 4.1.2 (Garmin International) and Garmin China City Navigator NT 2012 digital elevation model (DEM) and topographic maps at 10 m resolution to locate the approximate study areas. Freely available global digital elevation tiles at 20 m resolution were downloaded (Advanced Spaceborne Thermal Emission and Reflection Global DEM Version 2 http://asterweb.jpl.nasa.gov/gdem.asp) and used to delineate the accurate basin areas of reservoirs using hydrological functions in ArcGIS. Land cover data were downloaded from Finer Resolution Observation and Monitoring-Global Land Cover (http://data.ess.tsinghua.edu.cn/) using the Landsat Path list option. This dataset was the first 30 m resolution global land cover maps produced using Landsat Thematic Mapper and Enhanced Thematic Mapper Plus data (Gong et al. 2013). The land cover of each reservoir basin was obtained by clipping the corresponding land cover tile with the basin boundary of the reservoir. Percent crop and percent forest cover were summed relative to total land cover. Only in one reservoir (L) was the land cover data assessed as forest when the land cover map indicated water (land cover data was taken in January), we assumed this to be an error as land cover immediately adjacent was displayed as forested. Land use ranged from highly forested (e.g., F, 85.4% forest cover) to primarily agricultural (e.g., T, 46.8% crop cover; Fig. 1; Table 2). Water and zooplankton sampling were done from a small boat in the open water zone of the reservoir in front of the dam. Water was collected for total nitrogen (TN), TP, and chemical oxygen demand (COD) using a one liter Van Dorn water sampler at one meter below the water surface. Samples for TN, TP, and COD were dispensed into acid-cleaned Nalgene bottles and frozen until analysis. Nutrient analysis was carried out by the Shanghai Environmental Monitoring Center (Shanghai, China). TN concentrations were measured according to the alkaline potassium persulfate digestion method (State Environmental Protection of China 1990a) and TP according to the ammonium molybdenum spectrophotometric method with ultraviolet-visible (UV) spectrophotometry following a potassium persulfate digestion (State Environmental Protection of China 1990b). COD was determined using the permanganate index (ISO 8467:1986). Water samples for phytoplankton Chl a pigments were dispensed into opaque Nalgene bottles, and 500 mL were filtered in duplicate through 1.2 μm GF/C filters and stored frozen in tin foil. Chl a analyses were completed at McMaster University (Ontario, Canada) following previously published methods (Chow-Fraser 2006). Briefly, frozen filters were placed in 10 mL of 90% reagent-grade acetone and kept in the freezer for two hours prior to centrifugation. Chl a content was determined by measuring absorbance by spectrophotometry before and after acidification with hydrochloric acid to account for phaeophytin pigments. The phaeophytin corrected values were used in this study. Unfiltered surface water (one meter) samples for field blanks and THg and MeHg analysis were taken by use of trace metal clean techniques. These concentrations represent an important, but preliminary assessment: we placed most of our effort into quantifying the fish food web, and due to restrictions on access to reservoirs in the field, we were unable to characterize temporal and spatial variability in water Hg concentrations. Surface water MeHg concentrations measured here may be an underestimate if hypolimnetic conditions are favorable to methylation, as can be observed in reservoirs that stratify. Information on stratification was not available at the time of sampling, and this limits our ability to determine trends in MeHg concentrations in water with indicators of eutrophication. Measurements of BAF (see equation below) that use water MeHg concentrations in the denominator may also be overestimated. The decision to take unfiltered surface waters as done previously by others (Rolfhus et al. 2011) was largely due to lack of access to appropriate ultratrace filtration equipment. Prior to field sampling, Teflon bottles were acid cleaned with concentrated nitric acid, dilute hydrochloric acid (HCl; 10%) and stored in nitric acid (HNO3; 10%). Bottles were dipped below the surface and the "clean hands and dirty hands" technique (US EPA 1996) was used during sampling to minimize potential for contamination. All water samples for Hg analysis were frozen immediately in the field; those for MeHg analyses were preserved with pure HCl prior to freezing. Phytoplankton were collected by vertical tows through the water column, using a Wisconsin net (45 μm nylon mesh; nine tows at one meter); zooplankton were collected with a cone net (202 μm nylon mesh; three tows at 7–20 m, from one meter above the bottom to the surface, with care taken to avoid collecting sediment). Phytoplankton species composition is given elsewhere and was significantly correlated to TP concentrations (Razavi et al. 2014a). The size fraction for zooplankton was selected to allow comparison with previously published results (Chen and Folt 2005; published data extracted using DigitizeIt 2.0.4, http//www.digitizeit.de). Zooplankton samples for enumeration (i.e., zooplankton density) were preserved in 95% alcohol, diluted to a volume of 100 mL, then subsampled and counted (up to 250 individuals per replicate). Zooplankton biomass was calculated on three additional zooplankton tows following Chen and Folt (2005); samples were filtered onto preweighed Whatman filters and then dried at 60°C and weighed. Separate zooplankton samples were collected for THg and MeHg analysis; samples were rinsed with ultraclean water and frozen in Teflon bottles in the field (bottles were prerinsed in acid as described earlier for water samples). Zooplankton samples were freeze-dried (Stewart et al. 2008), homogenized, and split into subsamples for Hg and stable isotope analysis. We sampled fish from all trophic levels and targeted the same species at all sites. Stocked fish (i.e., raised in ponds and hatcheries to a certain length and released into the reservoir as fingerlings) and wild fish were sampled, but farmed fish were not included as those are exclusively fed with fish food in various types of enclosures, and thus, do not represent the reservoir-wide food web. Details of fish diet items are provided in Razavi et al. (2014b). With the guidance of local fisheries managers who were very familiar with both fish management and fish vendors at each reservoir, we sampled fish from individual fishermen and private fishing enterprises, as well as specific vendors that sold exclusively live fish from the reservoir at local markets. Individual fish weights and total length ranges were measured (Web Appendix, Table A1, www.aslo.org/lo/toc/vol_xx/issue_x/xxxxxa.xxx) and fish dorsal muscle tissues were removed in the field, transported on ice, and frozen at facilities at Fudan University prior to analyses in Canada. Tissues were then dried at a low temperature (between 50°C and 60°C) for 48 h for Hg and stable isotope analysis in a Hg-free oven (no Hg thermometers were ever used and the oven was periodically brought to 180°C to remove residual Hg) at Queen's University (Ontario, Canada). Dried fish tissues were required for the digestion method (see below) and oven drying caused the lowest Hg losses relative to other drying methods (Ortiz et al. 2002). Stable isotope analyses A portion of each dried muscle sample was homogenized finely using stainless steel jars and balls on a Mixer Mill MM200 (Retsch GmbH & Co KG), and then weighed into tin capsules (∼ 1.2 mg) using a calibrated semi-micro balance (Sartorius AG). Stable isotopes of nitrogen were analyzed at the Stable Isotope Facility at the University of California, Davis by a Sercon-Gas Solid Liquid elemental analyzer interfaced to a Sercon 20-20 isotope ratio mass spectrometer (Sercon). Samples were interspersed with blanks and standards (standard deviation, SD, δ15N 0.3‰); the standards were previously calibrated against National Institute of Standards and Technology (NIST) reference materials. Triplicates of in-house standard materials, defatted Atlantic salmon (mean δ15N = 8.8‰, n = 12) and farmed tilapia (mean δ15N = 6.1‰, n = 12) were included in each run; SD were less than 0.1‰. A duplicate of a fish sample was run every 10 samples as a check on consistency (coefficient of variation: δ15N = 0.8%). Delta values of nitrogen were calculated with the equation δ15N = ([(15N: 14N sample): (15N: 14N standard)] − 1) × 1000. The δ values are expressed in per mil (‰) deviation from international standards (i.e., Air). Mercury analyses Unfiltered water samples and zooplankton samples were analyzed for THg and MeHg by the Laboratory for the Analysis of Natural and Synthetic Environmental Toxins research facility at the University of Ottawa (Ontario, Canada). THg concentrations in water samples were analyzed by cold vapor atomic fluorescence spectrometry (CVAFS) using US Environmental Protection Agency (EPA) Method 1631 (US EPA 2002). The reporting limit (lowest point in calibration curve) for THg in water was 0.2 ng L−1. THg concentrations for freeze-dried zooplankton were measured using a direct thermal decomposition Hg analyzer, with a method detection limit of 0.012 ng g−1. Quality assurance for zooplankton included National Research Council of Canada (NRC-CNRC) certified reference materials (CRM) DORM-4 (fish protein; 92.7-101.2% recovery, n = 2), TORT-2 (lobster hepatopancreas; 115.6% recovery, n = 1) and the NIST standard reference material NIST-1566B (oyster tissue, 84.0-85.7% recovery, n = 2). Methylmercury in water was analyzed by capillary gas chromatography (GC) coupled with CVAFS (Cai et al. 1996). Zooplankton MeHg was extracted into dichloromethane and subsequently analyzed by GC CVAFS (Cai et al. 1997). Recovery of CRM for MeHg was 89.5-94.9% for DORM-4 (n = 2) and 97.4-98.1% for TORT-2 (n = 2). The method detection limit for MeHg in water and solids was 0.002 ng L−1 and 0.016 ng g−1, respectively. Recovery of MeHg in spiked water samples was 91.5% ± 2.7% (n = 4). Field blanks [0.2 ± 0.1 ng L−1, n = 4 for THg; 0.009 ng L−1 (n = 1) for MeHg] and procedural blanks revealed no contamination during sampling and analysis. All fish THg analyses were carried out at Queen's University using CVAFS (Tekran 2600 Total Hg Analyzer, Tekran Instruments, Ontario, Canada) following US EPA Appendix to Method 1631 (US EPA 2001a). All equipment and glassware used were washed in ultrapurified trace-metal quality water (EMD Millipore), placed in an acid bath for a minimum of 12 h, and rinsed again in Millipore water. Oven dried fish tissue (∼ 25 mg) was weighed into Teflon vessels, and digested at 200°C for 15 min in trace metal grade acid (5 m L−1 HNO3) and hydrogen peroxide (5 m L−1 H2O2) in a Microwave Accelerated Reaction System (CEM). Samples were diluted using Millipore water, preserved using 0.5% bromine monochloride, and refrigerated at 4°C for at least 24 h prior to analysis. Quality control was assessed by analysis of CRMs: DORM-3, TORT-2, and DOLT-4 (dogfish liver, NRC-CNRC). Recovery of THg for DORM-3 was 109.4% ± 6.9% (mean ± SD, n = 22), for TORT-2 was 113.1% ± 5.1% (n = 13), and DOLT-4 was 104.5% ± 4.6% (n = 12). The reporting limit was 0.05 ng L−1 and blank values were all below 0.4 ng. Samples (n = 33) spiked with 0.4 ng Hg resulted in a mean spike recovery (88%) and spike precision (13%) that met quality control acceptance criteria (US EPA 2001a). Statistical analyses Effect of eutrophication We used Chl a, zooplankton density, and TP as indicators of eutrophication, as reported commo
Referência(s)