The biogeochemical behaviour of selenium in two lentic environments in the Elk River Valley, British Columbia

The biogeochemical behaviour of selenium (Se) in two lentic environments (Goddard Marsh (GM) and Fording River Oxbow (FRO)) was assessed through detailed examination of Se speciation in bottom water, porewater and sediment components. The depositional environments at GM and FRO differ with regards to organic matter content, organic matter sources (as revealed by C:N ratios) and redox character. X-ray absorption near edge spectral (XANES) data suggest that elemental Se and organo-Se represent the dominant hosts for Se at GM and FRO. At both sites, the vertical distributions of dissolved Se species in porewater are closely linked to the profiles of redox-sensitive metabolites. Porewater profiles indicate that the sediments at GM and FRO are serving as diffusive sinks for Se through in situ adsorption/precipitation of Se in suboxic horizons. Although the sediments at both sites serve as net sinks for dissolved Se, interfacial peaks in dissolved selenite (Se) and organo-Se demonstrate these species are recycled back into the water column. The conditions present at GM are more favourable for the recycling of reduced Se species. Such observations can be linked to subtle differences in redox conditions as illustrated by profiles of redox-sensitive species (dissolved NO3, Fe, Mn, SO4 and ΣH2S). These differences have important implications to both the recycling of reduced Se species into the water column and Se uptake by aquatic biota. Implications with regards to Se management, bioremediation and biologically availability (food chain transport) are discussed. INTRODUCTION The fine-grained organic-rich substrates typical to lentic systems (e.g., wetlands, ponds, lakes) serve as optimum media for the microbially-mediated transformations of selenate (Se) to reduced forms, including selenite (Se), elemental selenium (Se) and organic species (Masscheleleyn and Patrick 1993; Zhang and Moore 1996; Simmons and Wallschlager 2005). Accordingly, understanding and quantifying mechanisms involved in Se cycling within lentic environments is required to assess the long-term fate of Se and risks to biological receptors. Studies to date conducted through the Elk Valley Selenium Task Force (EVSTF), including assessments of fish, waterbirds, waterfowl and amphibians (McDonald and Strosher 2000; Minnow 2004; Golder 2005), have advanced our understanding of the effects of Se on biological receptors in both lentic and lotic environments. However, there remains a dearth of information with respect to the biogeochemical mechanisms controlling the speciation, accumulation and remobilization of Se within lentic environments in the region. B.C.’32 Annual Mine Reclamation Symposium Technical Paper 8 2 To expand our current understanding of Se behaviour in lentic systems, studies were conducted in two lentic environments in the Elk River Valley of southeastern B.C. This study represents a collaborative effort between the EVSTF, Lorax Environmental Services Ltd., Trent University (Dr. Dirk Wallschlager), University of Saskatoon (Drs. Cheryl Wiramanaden and Ingrid Pickering) and Laurentian University (Dr. Nelson Belzile). The approach focused on the collection of high vertical-resolution profiles of Se species in sediment, bottom water and porewater, with the primary objective being to delineate the biogeochemical processes governing Se behaviour. The results have both local and global relevance to our understanding of Se behaviour in aquatic systems. The preliminary work presented herein, prepared for the EVSTF, has not been fully reviewed nor endorsed by its membership. FIELD AND ANALYTICAL METHODS Environmental Setting Field surveys at Goddard Marsh (GM) (Elkview Coal Operations) and Fording River Oxbow (FRO) (Fording River Operations) were conducted between August 21-23 and September 4-7, 2007. These lentic zones were selected for study based on previous work at these sites, ecological significance, and proximity to mine-related inputs. GM is located immediately downstream of a sediment-pond discharge from Elkview Mine, and comprises a dense cattail (Typha latifolia) marsh with limited areas of open water. Water depths range from ~0.5 to 1.0 m. Sediments at GM are organic rich and fine-grained. FRO is located adjacent to the Fording River ~9 km downstream of the Fording Mine. FRO extends for several hundred metres and comprises narrow channels and open ponds which are hydraulically connected to the Fording River. Water depths at FRO range from 0.5 to 1.5 m. The system is replete in organic matter and hosts fine-grained sediments. Field Methods Duplicate sediment cores were collected by hand from GM and FRO using 8 cm diameter butyrate tubing. Cores were extruded and sliced at intervals ranging from 1 cm in thickness near the sediment-water interface to 5 cm in thickness at deeper sediment depths. Sediment sub-samples were placed in polyethylene bags and frozen prior to transport. The post-depositional behaviour of Se and exchange with the overlying water column was assessed by sampling of the porewaters and bottom waters using dialysis arrays (peepers), as described in Martin et al. (2002, 2003). The peepers afford 7 mm-resolution profiling of dissolved constituents (0.45 μm pore size) from ~20 cm above the benthic boundary to a sub-interface depth of ~30 cm. Dissolved metal samples were acidified to pH <2 with ultrapure nitric acid while samples for nutrients and sulfate analysis were frozen. For hydrogen sulfide analysis, a 2.0 mL sample was taken and spiked with 50 μL of 1 M zinc acetate. Samples for Se speciation analysis were frozen with dry ice immediately upon collection. Analytical Methods Total Se in porewaters was determined by inductively-coupled-plasma dynamic-reaction-cell massspectrometry (ICP-DRC-MS). Inorganic Se species in porewater were determined by anion-exchange B.C.’32 Annual Mine Reclamation Symposium Technical Paper 8 3 chromatography coupled to ICP-DRC-MS (AEC-ICP-DRC-MS), similar to Wallschlager and Roehl (2001). Dissolved organic selenium was converted to selenite (Se) by selective UV-photo-oxidation, and then determined indirectly (by subtraction of the sample’s native selenite (Se) concentration) by hydride generation-atomic fluorescence spectrometry (HG-AFS) (Chen et al., 2005). Determinations of dissolved trace element concentrations were performed using inductively-coupled plasma mass spectrophotometry (ICP-MS) at the Pacific Centre for Isotopic and Geochemical Research, University of British Columbia. Sulfate and nitrate concentrations in porewaters were measured by ion chromatography and total sulfide (ΣH2S = S, HS and H2S) was measured spectrophotometrically. Total carbon and sulphur concentrations in sediments were determined by combustion/gas chromatography at the University of British Columbia. Carbonate carbon was determined by coulometry. Organic carbon was determined by subtracting carbonate carbon values from the total value. Trace elements were analyzed by inductively-coupled plasma optical emission (ICP-OES) and mass spectrometry (ICP-MS) using solutions prepared by fusing sub-samples in lithium metaborate (LiBO2), followed by dissolution of the quenched glass in 10% nitric acid (HNO3). X-ray absorption near edge spectra (XANES) data were collected using the synchrotron at the Canadian Light Source Saskatoon, SK. XANES probes the absorption characteristics of a particular electron shell using tunable synchrotron light. The geometry of the resulting spectra is valence dependent so it is possible to determine the specific elemental oxidation states present (i.e., Se, Se, Se, Se) (Pickering et al., 1995). As well, XANES spectra can be used to obtain semi-quantitative determinations of the relative abundance of each oxidation state. To quantify the relative contribution of various Se forms, a XANES library of known Se compounds was compared to the sample spectra. RESULTS AND DISCUSSION Sediments Contrasts between the depositional environments at GM and FRO are illustrated by their carbon (C), nitrogen (N) and sulfur (S) content (Figure 1). The greater organic carbon content at GM (25 to 30 wt.%) in comparison to FRO (5 to 7 wt.%) likely relates to differences in the source(s) of organic matter. Specifically, the C-Org:N ratio in GM sediments (mean = 38) is closer to the C:N signature of terrestrial organic matter (45 to 50:1), while the C-Org:N ratio at FRO (mean = 20) is similar to organic matter produced by plankton decomposition (12:1) (Wetzel 1975). The higher C:N at GM indicates that the organic matter content at this site is composed largely of decomposing wetland vascular plants (e.g., Typha sp.). Conversely, the lower C:N contents at FRO imply a greater portion of the organic matter originates from in situ sources such as algal production. These differences likely have relevance to Se accumulation in sediments. Total-Se in sediments at GM range from 7 to 71 mg/kg dry wt. (mean = 37 mg/kg) while deposits at FRO range from 2 to 19 mg/kg dry wt. (mean = 10 mg/kg) (Figure 2). XANES spectra suggest that elemental Se, organo-Se (possibly seleno-methionine) and selenite (sorbed Se), are the dominant hosts for Se at B.C.’32 Annual Mine Reclamation Symposium Technical Paper 8 4 both GM and FRO (Figure 2). Of these, elemental phases and organo-Se contribute most to the total sediment inventory, which account for on average 35% and 50%, respectively, of the total at both sites. Org-C (wt.%) 0 5 10 15 20 25 30 35 0