Sandimirov S. Heavy metal contents in whitefish (Coregonus lavaretus) along a pollution gradient in a subarctic watercourse. Environ. Monit. Assess. 2011, V. 182, №1-4, p. 301-316.

Environ Monit Assess (2011) 182:301-316 305 Metal analyses Heavy metal analyses of water, sediment and fish tissue included nickel (Ni), copper (Cu), cadmium (Cd), zinc (Zn), lead (Pb) and mercury (Hg). These elements have in previous studies shown elevated levels in sediments in lake localities in the vicinity of the Nikel smelters (Dauvalter 1994, 1998; Dauvalter and Rognerud 2001). All analyses were conducted in the analytical laboratory of the Institute of North Industrial Ecology Prob­ lems, Kola Science Centre, Russian Academy of Sciences (Accreditation certificate No POCC RU.0001.517126). Water samples were taken at the depths of 0.1-0.3 m from the surface. pH was measured immediately during sampling. Samples for deter­ mination of metal contents were filtered through a 0.45-^m membrane filter, preserved with nitric acid and transported to the laboratory, where the metal concentrations were determined by flame and flameless atomic absorption spectrophotom­ etry. Ni, Cu and Zn were determined with Perkin Elmer Model 5000 furnace with graphite sprayer HGA-400, Cd and Pb with Perkin-Elmer 5100 with graphite sprayer P-E AAnalist800 and Hg with the mercury analysis system RMS-100 (Perkin-Elmer). Sediment samples were collected with a grav­ ity sediment corer, and vertically extruded and sectioned in 1 cm layers for analyses (see Dauvalter 1994; Dauvalter and Rognerud 2001). For the metal analyses, 0.2 g sediment sample (dry weight) was taken and digested in a Teflon ‘bomb’ for 4 h at 140°C with 2 ml concentrated nitric acid. Concentrations of Ni, Cu, Cd, Zn and Pb were analysed by atomic absorption spectrophotometer (Perkin Elmer 460 and 560) using the standard addition technique. Hg was determined utilizing cold vapor atomic absorption. Anthropogenic im­ pacts on the water ecosystems was assessed by calculating the contamination factor ( C f) for each investigated heavy metal, where C f values were determined as the ratio of the heavy metal concen­ tration in the surficial 1-cm layer divided by the background concentration in the deepest layer of the sediment core (Hakanson 1980, 1984). Fish tissue samples were dried to constant weight in a drying chamber at 105°C. The samples were further digested according to the manual of fish tissue decomposition in a mixture of nitric acid (4 ml), hydrochloric acid (0.5 ml) and wa­ ter (2 ml) in Teflon autoclaves in a microwave oven (Multiwave 3000, Anton Paar, Austria) at 600 watt for 13 min and at 400 watt for 10 min. After 20 min cooling, the solution was diluted to 15 ml with de-ionized water. The determination of Ni, Cu and Zn was carried out by atomic absorp­ tion spectroscopy (AAS) using a Perkin-Elmer Model 5000 furnace with graphite sprayer HGA- 400 after dilution of the initial sample from five to 50 times. The determination of Cd and Pb was carried out by AAS using a Perkin-Elmer 5100 with graphite sprayer after dilution from five to 500 times depending on the sample. Hg was de­ termined with the mercury analysis system FIMS- 100 (Perkin-Elmer) with sample dilutions from five to 500 times depending on the tissue. SnCl was used as reducing agent. The concentrations of metals in fish are expressed as micrograms per gram of dry weight tissue, with detection limits of 0.05 |xg g-1 for Ni, 0.02 |xg g-1 for Cu, 0.01 |xg g-1 for Cd, 0.05 |xg g-1 for Pb and 0.005 |xg g-1 for Hg. A quality assurance program was carried out for the heavy metal analyses using the standard ref­ erence material DORM-2 (dogfish muscle; NRC Institute for National Measurement Standards, Canada). Statistical analyses Dependence of fish size and age for the concen­ trations of the investigated elements were mod­ eled using linear mixed-effect modeling (hereafter referred as LME), using morph + fish weight (or age) as fixed effects, and lake locality as a random intercept effect (Pinheiro et al. 2009). Element concentrations were log-transformed and model residuals inspected. The possible effect of weight and age were modeled separately, as these are highly confounded variables. Differences in metal contents between lake localities and whitefish morphs was tested using Kruskal-Wallis test or Mann-Whitney U test with routine Bonferroni adjustments. The levels of metal content in fish tissues were compared to the levels of the cor­ responding metal in water samples and surficial sediment samples by Pearson correlation test. Springer