British Columbia Mine Reclamation Symposium

Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment… Haakensen, M.; Pittet, V.; Spencer, J.; Rodgers Jr., J. H.; Castle, J. W. 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


59367-Haakensen_M_et_al_Process_driven_design_and_piloting_Yukon.pdf [ 1.16MB ]
JSON: 59367-1.0305886.json
JSON-LD: 59367-1.0305886-ld.json
RDF/XML (Pretty): 59367-1.0305886-rdf.xml
RDF/JSON: 59367-1.0305886-rdf.json
Turtle: 59367-1.0305886-turtle.txt
N-Triples: 59367-1.0305886-rdf-ntriples.txt
Original Record: 59367-1.0305886-source.json
Full Text

Full Text

Mine Closure 2015 – A.B. Fourie, M. Tibbett, L. Sawatsky and D. van Zyl (eds) © 2015 InfoMine Inc., Canada, 978-0-9917905-9-3 Mine Closure 2015, Vancouver, Canada 545 M. Haakensen  Contango Strategies, Canada  V. Pittet  Contango Strategies, Canada J. Spencer  Capstone Mining Corporation, Canada J.H. Rodgers Jr.  Clemson University, USA J.W. Castle  Clemson University, USA   When designed and executed in a scientifically guided manner, constructed wetland treatment systems (CWTSs) can treat various contaminants in water. A process-driven treatment plan was developed to convert contaminants in the water to minerals in the soil, improving water quality and decreasing risk to the environment. This site-specific CWTS is being developed for Capstone Mining Corporation’s Minto Mine (Yukon), using a scaled phased approach to allow for improvement, optimisation and flexibility for modifications along each step. These phases are: (1) site assessment and information gathering, (2) technology selection and conceptual design, (3) pilot-scale testing and optimisation (controlled environment), (4) onsite demonstration-scale confirmation and optimisation and (5) full-scale implementation. The first three phases have been completed successfully, with Phase 4 underway. Performance highlights include successful application of microbial profiling technologies (genetic and growth based) to guide system design in a site-specific context. Microbiome technologies applied in a site assessment and throughout pilot-scale testing allowed for identification of natural copper- and selenium-attenuating microbial communities and ecosystems at the mine site, correlation with native wetland plant species and confirmation of the enhancement of beneficial microbial populations through design of the CWTS. The pilot-scale CWTSs confirmed plant amenability to transplantation, and the design selected for further testing on site achieved 92% removal of copper (mean influent 146 μg/L, outflow 11.3 μg/L) and 41% removal of selenium (mean influent 10.2 μg/L, outflow 6 μg/L) using synthetic influent designed to mimic the worst-case water chemistry of a long-term closure scenario. A mass-balance of the pilot-scale systems confirmed that the elements were sequestered to sediments of the CWTS, with less than 0.5% of the copper and 2% of the selenium transferred to the plant leaves. The pilot-scale system allowed for selection of the optimal design (from three different designs) and for different water chemistries to be tested for treatment, mimicking early closure (containing ammonia and nitrate) and long-term closure (no longer containing ammonia and nitrate). At the conclusion of testing, the pilot-scale CWTS was converted to a hybrid bioreactor/CWTS to evaluate its potential to function as a semi-passive treatment system, should the need arise (e.g. change in regulatory requirements or change of influent water chemistry). Conversion to the hybrid bioreactor/CWTS allowed identification of critical aspects of the transition period and improvement of selenium removal, having influent of 12 μg/L and outflow of 3.9 μg/L, with lowest recorded outflow concentration of 1.9 μg/L (84% removal). Data from the pilot-scale testing were used to calculate system-specific removal rate coefficients, allowing for more accurate sizing estimates for full-scale and modelling effects of seasonal flow-rate and water chemistry variations. The optimised design was applied to construct a demonstration-scale CWTS at the Minto Mine during fall 2014 to refine findings and calculations, including seasonal variations.   Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 546 Mine Closure 2015, Vancouver, Canada The Minto Mine, operated by Capstone, is located 240 km northwest of Whitehorse, Canada, on the west side of the Yukon River. The Minto property lies within the eastern part of the Dawson Range, with elevations from 700 to 1,000 m; the landscape has rounded mountains intersected by broad valleys and drainages that are part of the Yukon River watershed. The climate in the Minto region is subarctic continental characterised by long, cold winters and short, cool summers. The area experiences moderate precipitation in the form of rain and snow and a large range of temperatures on a yearly basis with a mean annual temperature below 0°C. The Minto Mine has been in commercial operation since October 2007, and the deposits being mined are copper sulphide mineralised zones. Surface water and groundwater quality is a key consideration in the evaluation of potential effects of mining and mineral development projects, and changes to water quality parameters have the potential to affect aquatic and human use of water resources. A reclamation and closure plan (RCP) is required under both the Yukon Territorial Waters Act and Waters Regulation in the form of a water licence and the Quartz Mining Act, under direction of the director of mineral resources, in the form of a quartz mining licence. The RCP is intended to address the long-term physical and chemical stability of the site and closure of the proposed features and disturbances associated with the mine. As part of the RCP, a constructed wetland treatment system (CWTS) is being designed, evaluated and optimised for water treatment at closure through a phased program. These phases are: (1) site assessment and information gathering, (2) technology selection and conceptual design, (3) pilot-scale testing and optimisation (controlled environment), (4) onsite demonstration-scale confirmation and optimisation and (5) full-scale implementation. The CWTS is being designed with the objective of treating water during free-flowing months. Low flow volumes in the winter will freeze and will discharge with snow melt during freshet. The findings from Phase 3 pilot-scale testing and optimisation in a controlled environment are summarised below. When designed and executed scientifically, CWTSs can treat various contaminants in water (Hawkins et al., 1997; Huddleston and Rodgers, 2008; Murray-Gulde et al., 2008; Rodgers and Castle, 2008). Once established, CWTSs can become self-sustaining ecosystems with the plants providing yearly renewal of carbon to fuel microbial activity. As such, they possess the desirable potential to remediate contaminated mine drainage for as long as it is generated. Having a site-specific approach allows for the appropriate set of operational conditions to be achieved, where complex coupled reactions can take place for treatment of targeted constituents. CWTSs have been used to mitigate risks to aquatic receiving environments from a variety of aqueous contaminants, including copper (Cu) and selenium (Se) (Kadlec and Wallace, 2009), which are of interest to the Minto Mine.  Many biogeochemical reactions take place in a wetland environment, but specific reactions may decrease the aqueous concentrations and bioavailability of some contaminants (e.g. Cu, Se), thus lowering risk to the receiving environmental systems by decreasing exposure to contaminants. Removal mechanisms include filtration and settling, sorption to soils or dead organic material, uptake into plants or biofilms, complexation, degradation, mineralisation and volatilisation. A process called dissimilatory sulphate reduction can chemically reduce available aqueous sulphate to sulphide through the metabolism of specific bacteria. Once dissimilatory sulphate reduction has occurred, the sulphide provides a mechanism to remove dissolved cations such as Cu from water by complexation to produce copper sulphide, an insoluble precipitate. The mechanism for selenium treatment in a wetland is more direct, through dissimilatory selenium reduction that is also microbially mediated, sometimes by some of the same bacteria that catalyse dissimilatory sulphate reduction. The most common forms of aqueous Se are selenite and selenate, with valence states of IV and VI, respectively. When these are reduced in a wetland to an insoluble state, Se precipitates as Case Studies Mine Closure 2015, Vancouver, Canada 547 elemental Se(0); however, at lower reduction states, it can also be reduced to selenide (-II) valence state, which will volatilise.  The insoluble forms of Cu and Se are then removed from the water through precipitation/settling and filtration mechanisms inherent in wetlands. These are then sequestered into the sediments over time through accretive processes. Accretion is the naturally occurring process of wetland sedimentary material (precipitates, decaying plant material, etc.) accumulation over time. CWTSs can be designed to have different accretion rates. Once an accreting CWTS is established and mature, targeted constituents are essentially sealed away under layers of sediment, rendering them less bioavailable and less susceptible to re-entry into the water column. Because of this design attribute, there is no need to dredge the CWTS or harvest wetland plants for the system designed for the Minto site; in fact, this type of activity would disrupt treatment functions and re-expose previously sequestered constituents. This process mimics what occurs in natural wetlands; therefore, it is often the best option for long-term, low-maintenance, effective treatment.  In order to fully understand the function of biogeochemical processes such as the mineralisation of aqueous dissolved copper and selenium, it is critical that the microbial populations performing these functions be well characterised and understood. For example, information about the total quantity and relative abundance of sulphate-reducing bacteria (SRB) can help gauge the effectiveness of a CWTS design and probable robustness of the system. Microbiome analyses and growth-based assays can provide insights into microbial communities in CWTSs, allowing for more effective decisions on design and optimisation while helping to determine the stability of microbial communities in fluctuating conditions. In order for CWTSs to be predictable and robust, they must be designed, piloted and optimised through phases in a site-specific manner to allow for improvement along each step (Haakensen et al., 2015). This manuscript describes results from pilot-scale testing that involved selecting the optimal CWTS design (from three different designs) and testing different water chemistries for treatment, mimicking early closure (containing ammonia and nitrate) and long-term closure (no longer containing ammonia and nitrate). At the conclusion of testing, the pilot-scale CWTS was converted to a hybrid bioreactor/CWTS to evaluate its potential to function as a semi-passive treatment system, should the need arise (e.g. change in regulatory requirements or influent water chemistry). This included the addition of an organic carbon amendment (alfalfa hay and straw) to stimulate reducing conditions for improved selenium treatment. Alfalfa hay and oat straw were chosen, as they would provide both short- and long-term carbon sources compared to a metred dosing of more available electron donors such as ethanol. Results from pilot-scale testing were then used to calculate system-specific removal rate coefficients, allowing for more accurate sizing estimates of the full-scale CWTS and modelling for effects of seasonal flow-rate and water chemistry variations.  Two plant species were of interest for the pilot-scale CWTSs based on data collected during site assessments (reports 2013-0100-256 and 2013-0100-257 on the YESAB registry). Carex aquatilis (sedge) was chosen, as it is readily available at the Minto site and was one of the first plants to colonise cleared or disturbed areas that were wet or flooded, indicating that it may be a good candidate for quickly establishing the CWTS. Furthermore, data from the site assessment suggested that Carex hosted a range of beneficial microbes and was conducive to reducing conditions, presumably with low radial oxygen loss (ROL). Moss was used in the second and third pilot-scale CWTS designs, as Carex and moss coexisted at Minto locations having shallower water. As such, we wanted to assess whether moss contributes positively or negatively to treatment. It is well known that moss has a high uptake rate of cations (such as copper and selenium) and is a relatively benign sink for these elements (i.e. is not a food source for invertebrates or higher animals, and as such does not contribute greatly to bioaccumulation [Haines and Renwick, 2009; Longton, 1997; Suren and Winterbourn, 1991]). This was confirmed with the site assessment in 2013, which produced data indicating that moss at the Minto Mine has a high uptake of elements such as copper, and could therefore be an effective means of removing elements from seepage through uptake.  Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 548 Mine Closure 2015, Vancouver, Canada Key considerations for plant selection for CWTSs include physiology (e.g. ROL, water depth tolerance), biomass production per year versus decomposition rate (to allow for accretion), bioconcentration tendencies, plant density in the wetland cell (to aid in flow distribution), flow rate tolerance, structural capacity and prevention of sediment re-suspension, evapotranspiration rate and provision for microbial habitat. ROL is a crucial parameter to consider when designing a wetland cell’s targeted conditions (e.g. oxidising vs. reducing; Taylor, 2009), which in turn directly affects treatment mechanisms (e.g. dissimilatory reduction). Furthermore, plant selection was based on site-specific information, as plant characteristics vary not only between plant species, but also between ecotypes, and it was important to plan for a CWTS design that would not be subjected to a drastic transition of plant species over time.  The pilot-scale CWTSs were designed using a process-driven approach based on information gathered through a site assessment specific to the purpose of guiding CWTS development or other applicable passive treatment technologies (reports 2013-0100-256 and 2013-0100-257 on the YESAB registry).  Pilot-scale surface-flow CWTSs were constructed in dedicated indoor greenhouse pilot facilities in Saskatoon, Saskatchewan, Canada. The ambient temperature of the indoor facilities was kept at 20°C (+/- 6°C) with minimum 12-hour photo periods. The CWTSs were planted with either Carex aquatilis (sedge), or Carex and aquatic moss (with and without 5% v/v pine biochar added to the soils) for a total of three types of CWTSs being tested in duplicate, resulting in six pilot-scale treatment systems in this study. Each system had two cells in a series in order to better define removal rate coefficients through a range of concentrations. Each cell had a 41 cm diameter and a height of 57 cm (Figure 1). Pine biochar was chosen as a soil amendment, as previous studies have demonstrated that addition of biochar to the hydrosoil can result in faster establishment and growth of newly planted wetland plants (Haakensen et al., 2013). Additionally, it was hypothesised that the biochar may provide greater surface area for beneficial microbial communities to establish. The hydrosoil mixture was the same for all pilots: sand amended with 2% woodchips (v/v) and 5% peat, except that the third system design also included a 5% pine biochar amendment. The sand and amendments were tested for metal and nutrient contents. The selected biochar was low in soluble salts and had a high surface area, low metals content and moderate available organic carbon and phosphorous. Moss and Carex were harvested from the Minto site and used to vegetate all pilot-scale CWTSs. To establish the plants, soil was filled to a depth of 20 cm at the bottom of each cell, and submerged under 10 cm of tap water. Ten to 12 Carex were planted per cell, with an approximately 250 mL volume of moss (compressed), as appropriate for the system design. The water depth was changed throughout the pilot-scale testing to evaluate effects on performance and develop strategies for optimisation.   Case Studies Mine Closure 2015, Vancouver, Canada 549 A thorough water characterisation was performed to determine the range of the following constituents: acidity, alkalinity, biological oxygen demand, carbonate, chemical oxygen demand, chloride, conductivity, hardness, ion balance, iron, manganese, metals, metalloids, pH, sulphate, total Kjeldahl nitrogen, total organic carbon and total suspended solids. This is particularly important when considering complex systems such as mining-influenced waters, as some constituents (e.g. chloride) can interfere with wetland function, while others can affect the removal of targeted constituents in beneficial or detrimental ways (e.g. iron, sulphate, ammonia, nitrate, alkalinity). zz The CWTSs were designed for treatment of the water chemistry predicted to be present at the Minto site after closure. As such, this water does not yet exist and had to be synthesised as an approximation based on water quality modelling. The worst-case post-closure water chemistry for the month of October was used, as it had the highest concentrations of key elements for any month expected to have free-flowing water. The simulated water recipe underwent five iterations of design, testing and optimisation to ensure the composition and method of assembling resulted in analytical chemistry that matched the predicted water quality. The recipe was then assembled with both high-nitrogen and low-nitrogen additions to mimic concentrations of blasting residues (ammonia and nitrate) in short-term and long-term closure scenarios. The final recipe and resultant water chemistry are provided in Table 1. Before performance testing began, the pilot-scale systems were established with tap water for approximately six weeks. This was followed by a 10-week acclimation period (i.e. no testing) with the high-nitrogen synthetic influent to avoid misconstrued data regarding treatment due to sorption, as well as to establish plants.  The pilot-scale systems were operated for more than 30 weeks, with testing in three phases to determine the effect of ammonia and nitrate concentrations and added organic carbon on constituent removal. The first phase of testing was over a six-week period to mimic short-term closure with a high nitrogen concentration (high N period) based on the predicted maximum worst-case ammonia and nitrate concentrations (Table 1). This was followed by a second phase of testing over approximately six weeks during which only very low concentrations of nitrogen (low N period) were present in the influent water in order to simulate a long-term closure scenario. Although the chemistry of the water is expected to improve over time in long-term closure (e.g. because of covers and establishment of vegetation), the chemistry was kept consistent for both periods of testing other than for nitrate (and accordingly ammonia, which can be oxidised to nitrate). This is because nitrate can affect the treatability of compounds, such as selenium, that are removed from water through coupled biogeochemical reactions and dissimilatory reduction, as nitrate has a high affinity for available electrons and is preferentially reduced before other compounds. However, it is known that after blasting activities have ceased, the concentrations of ammonia and nitrate will decrease over time. If other constituents were modified at the same time, it would not be possible to ascertain the effects of nitrate on performance. Therefore, the two periods (high and low N) were tested to gain information about the expected performance of the system under these two scenarios.     Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 550 Mine Closure 2015, Vancouver, Canada Element (total) Unit Targeted concentration1  Average influent concentration by testing period Acclimation  High N Low N Hybrid bioreactor-CWTS Aluminium (Al) mg/L 0.100 0.089 0.085 0.080 0.098 Arsenic (As) mg/L 0.0030 0.0031 0.0033 0.0033 0.0037 Cadmium (Cd) μg/L 0.334 0.011 0.236 0.336 0.392 Chromium (Cr) mg/L 0.0065 0.0058 0.0064 0.0071 0.0075 Cobalt (Co) mg/L 0.0039 0.0039 0.0042 0.0043 0.0044 Copper (Cu) mg/L 0.13 0.140 0.146 0.146 0.176 Iron (Fe) mg/L 3.1 3.1 3.1 3.2 3.5 Lead (Pb) μg/L 0.70 0.44 0.67 0.59 0.71 Magnesium (Mg) mg/L 50 49 49 51 53 Manganese (Mn) mg/L 2.6 2.5 2.5 2.6 2.7 Molybdenum (Mo) mg/L 0.012 0.013 0.015 0.013 0.013 Nickel (Ni) mg/L 0.0072 0.0077 0.0087 0.0076 0.0080 Selenium (Se) mg/L 0.010 0.010 0.011 0.010 0.011 Sodium (Na) mg/L 44 40 41 42 43 Strontium (Sr) mg/L 2.9 2.5 2.5 2.5 2.6 Sulphur (S) mg/L 85 85 84 86 88 Uranium (U) mg/L 0.0051 0.0012 0.0012 0.0012 0.0014 Vanadium (V) mg/L 0.016 0.015 0.013 0.014 0.018 Zinc (Zn) mg/L 0.032 0.052 0.037 0.040 0.042 Other Parameters       Total ammonia (N) mg/L 0.2-0.5 2 0.61 0.50 0.29 0.05 Dissolved nitrate (N) mg/L 0.8-28 2 29 30 0.86 0.84 pH pH 7.5-8.0 7.89 8.16 7.96 8.01 Alkalinity (Total as CaCO3) mg/L - 107 107 100 100 Bicarbonate (HCO3) mg/L - 133 130 122 130 Water depth cm - 10 10 17 3 17 30 Hydraulic retention time hr - 37± 3 37± 3 53 ± 3 3 53 ± 2.5 87 ± 3 1 Targeted concentration based on modelled worst-case post-closure predicted water quality. 2 Ammonia and nitrate were not added to the synthetic water for the low N and hybrid bioreactor-CWTS periods. 3 Water depth was changed from 10 cm to 17 cm during the high N period to test plant tolerance to deeper water.   The third and final stage of pilot-scale testing included the addition of an organic carbon amendment (alfalfa hay and straw) to stimulate reducing conditions for improved selenium treatment. This hybrid bioreactor-CWTS period was conducted for approximately seven weeks to test how the CWTS functions once converted from conventional CWTS operation to the hybrid configuration. During the hybrid bioreactor-CWTS testing period, 1.4 g of alfalfa pellets were added per litre of water in each of the first cells in a series. Straw was added at 10 cm3 per L to the first cells, and 5 cm3 per L to the second cells in a series. The alfalfa pellets and Case Studies Mine Closure 2015, Vancouver, Canada 551 oat straw were subjected to analytical and leachability testing prior to use to ensure suitability for this application.  Constant flow metering pumps were used to set the flow rate of water into the system. Flow rates were confirmed regularly (at minimum weekly) and adjusted as necessary to within +/−10% of the desired flow, to achieve a known hydraulic retention time (HRT). Outflow rates were also measured to confirm that they were not significantly different from inflow rates, indicating that evaporation was not a major influence on water quality (i.e. through concentration).  Dissolved oxygen (DO; mg/L), temperature (°C), pH, specific conductivity (SPC; μS/cm), and oxidation-reduction potential (ORP; mV) were measured on a routine schedule with an YSI Professional Plus handheld unit. The relative oxidation-reduction potential of the soil (referred to in this document as ‘relative redox potential’ to avoid confusion with water ORP measurement) was measured using inert electrodes (copper wire probes with platinum tips) that were permanently installed in the soil of the cells and remained there for the duration of the project to ensure accurate readings (Faulkner et al., 1989). To take a reading of the relative redox potential between the soil and water, a reference electrode (Accumet Calomel) was suspended in the water column above the inert electrode and measured in millivolts by a voltmeter. Water, sediment and plant samples were collected throughout the pilot-scale study, with analytical testing performed by a third-party accredited laboratory. Outflow concentrations for pilot-scale system designs and/or testing periods were compared, with a p-value calculated using the non-parametric Kruskal-Wallis test. Sediment samples were also collected for the microbial analyses outlined in Section 2.5. In addition to routine analysis for metals in plants, a replicate of each plant sample (Carex and moss) collected at the end of the low N period was used in a trial to determine whether the plants leach constituents of concern, such as Cu and Se, after freezing and thawing of surrounding water. In this trial, 200 mL of water from the corresponding CWTS cell was added to the plant sample to completely cover the plant material. The samples were frozen at −20°C for five days, and then thawed at room temperature (~22°C) overnight. The plant samples were removed and lightly squeezed to return residual water to the container, and the plant tissue was sent for metals analysis to compare with pre-freeze concentrations.  A removal rate coefficient (k) can be calculated according to Equation (1).   𝑘 =−ln⁡(𝐶𝑓 𝐶𝑖)⁄𝑡 (1) Cf  =  final concentration  Ci  =  initial concentration  t  =  HRT  Once k has been calculated from known data (such as the data used in this study), the equation can be rearranged to solve for the HRT (t) and, hence, estimated full size needed for the CWTS to achieve outflow objectives, given a known influent concentration and flow rate. Likewise, it can be used to solve for an outflow concentration, given a known retention time and influent concentration, thereby being a necessary tool for use in any CWTS design. That said, removal rate coefficients are highly specific and must be developed in a site-specific manner for each element of interest. While they may sometimes be applied in a conceptual manner to other situations/sites, caution should be taken in applying a removal rate coefficient developed for one design and water chemistry to a very different chemistry or design basis. It is often the case (as it is here for copper) that k must be calculated and applied for different ranges of certain constituents. For this reason, the pilot-scale CWTSs were constructed with two cells in a series to test different ranges of element concentrations as they are treated through the system. Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 552 Mine Closure 2015, Vancouver, Canada Most-probable number (MPN) assays were performed for sulphate-reducing (active standard ASTM D4412), selenite-reducing (Siddique et al., 2006) and nitrate-reducing (nitrate reduction test, supplied by Sigma-Aldrich) organisms. Total heterotrophs (grown with YTS250 medium; Lefrançois et al., 2010) were also quantified in aerobic and anaerobic conditions. In brief, sediment samples were diluted 1:100 with a 0.1% peptone solution. This starting dilution was then diluted serially from 1/400 to 1/419,430,400. At minimum, all tests were conducted in duplicate. Wells were incubated at +30°C without light and assessed for visible growth (formation of a bacterial pellet) and/or colour change specific to the type of media according to the publication or manufacturer’s protocol after 27 days. The MPN of organisms capable of each metabolism was then calculated as previously described (Blodgett, 2010). The MPN of organisms in each pilot-scale system was compared over time using the non-parametric Kruskal-Wallis test. DNA was extracted from all sediment samples using the MO BIO Powersoil Powerlyzer DNA extraction kit, with the addition of phenol during cell lysis per the manufacturer’s alternative protocol. Targeted DNA sequencing was used to identify bacteria present in each sample via polymerase chain reaction (PCR) amplification of the v3/v4 region of the 16S ribosomal RNA gene (Klindworth et al., 2013). Library preparation and sequencing were performed per the manufacturer’s instructions for MiSeq v3 paired-end 300 bp sequencing (Illumina) for all samples and positive and negative controls. All raw sequences were filtered to remove low-quality reads based on the following criteria: average quality less than Q30, shorter than 350 bp, not having an exact match to the forward primer and having any base called N (unknown). Bioinformatics pipelines consisting of internally developed scripts and selected QIIME scripts (Caporaso et al., 2010; Edgar, 2010) were used to process the reads. Similar sequences were clustered into groups called operational taxonomic units (OTUs) using a 97% identity threshold and the script. All OTUs with less than 10 representative sequences across all samples were discarded as a quality-filtering step to remove OTUs that may have arisen owing to sequencing errors. Taxonomic classification of the OTUs was performed using the Greengenes database version 13_8 (DeSantis et al., 2006; McDonald et al., 2012). The percentage of the microbial community classified as organisms of interest (e.g. SRB) was also compared across samples. Weighted UniFrac distance measure (Lozupone et al., 2011) was used for ordination analyses using the “phyloseq” package (McMurdie and Holmes, 2013) in R. The most abundant OTUs were compared to the Cu, Se and S concentrations in the sediment to calculate the Spearman correlation coefficient and associated p-value. Over the course of the study, a total of 45,000 L of synthetic water (influent) was received by the pilot-scale CWTSs (in 41 batches), approximately 1,200 L of synthetic water per week, in a pilot-scale CWTS footprint of 1.56 m2. The system operated with a flow rate of 20 mL/min, resulting in an HRT ranging from 34 to 94.3 hours, depending on water depth during the respective period of testing (Table 1).  The performance of each pilot system, based on minimum, maximum and average measurements of copper and selenium concentrations as well as several parameters that aid in understanding treatment performance, is summarised in Table 2 for all three testing periods. The Carex and moss systems achieved an average copper removal of 92% (mean influent 0.146 mg/L, mean outflow 0.0113 mg/L) and average decrease of 41% for selenium (mean influent 0.0102 mg/L, outflow 0.006 mg/L) using synthetic influent designed to mimic worst-case water chemistry for a long-term closure scenario (i.e. low N period). While copper treatment was relatively stable through the high N and low N periods, selenium treatment took longer to establish and improved though the high N period, and then was stable during the low N period (Figure 2). While biochar Case Studies Mine Closure 2015, Vancouver, Canada 553 had no significant effects on performance, for both the high N and low N periods, all systems planted with moss and Carex had significantly better copper treatment than systems with Carex only (p < 0.005). In contrast, there was no statistically significant difference among the three pilot-scale system designs for selenium removal in any of the testing periods.  Period  Copper (mg/L) Selenium (mg/L) pH DO (mg/L) ORP  (water, mV) High nitrogen Influent Avg 0.146 0.0102 8.00 - - Carex Avg 0.018 0.0076 7.49 5.36 157 Min 0.016 0.0058 6.99 4.08 82 Max 0.020 0.0097 7.94 6.69 215 Carex + moss Avg 0.013 0.0073 7.64 5.35 247 Min 0.011 0.006 7.19 2.59 104 Max 0.016 0.0095 8.29 7.55 215 Carex + moss + biochar Avg 0.012 0.0067 7.84 5.75 140 Min 0.009 0.0058 7.39 3.50 84 Max 0.014 0.0085 8.28 8.70 205 Low nitrogen Influent Avg 0.146 0.0102 7.96 - - Carex Avg 0.021 0.0061 7.54 5.61 168 Min 0.017 0.0053 7.16 3.99 126 Max 0.032 0.0067 8.05 7.72 220 Carex + moss Avg 0.011 0.0060 7.71 5.73 159 Min 0.010 0.0056 7.26 3.02 113 Max 0.013 0.0065 8.32 9.78 188 Carex + moss + biochar Avg 0.012 0.0063 7.6 5.21 182 Min 0.011 0.0058 7.15 2.57 106 Max 0.015 0.0069 8.25 9.04 374 Hybrid bioreactor-CWTS Influent Avg 0.176 0.0119 8.01 - - Carex Avg 0.021 0.0040 7.16 1.52 -129 Min 0.0075 0.0019 6.91 0.02 -175 Max 0.045 0.0056 7.62 3.23 -79 Carex + moss Avg 0.023 0.0040 7.20 1.73 -150 Min 0.0083 0.0025 6.95 0.11 -212 Max 0.044 0.0054 7.82 5.61 -101 Carex + moss + biochar Avg 0.026 0.0040 7.19 1.86 -134 Min 0.0083 0.0028 6.96 0.03 -195 Max 0.064 0.0054 7.71 4.71 -30 * The average outflow for the hybrid bioreactor-CWTS period includes data from 19 June, when there was a spike in concentrations. If this data point is not included, the average outflow concentrations are 0.015, 0.017 and 0.015 for Cu mg/L, and 0.003, 0.003 and 0.004 for Se mg/L for the Carex, Carex + moss and Carex + moss + biochar systems, respectively. The hybrid bioreactor-CWTS configuration was tested as a contingency measure in case higher removal rates are needed than achieved by the CWTS at a given point in time (e.g. influent water quality or outflow objectives change). There was no significant difference between the treatment performances of the three system designs for either copper or selenium during the hybrid bioreactor-CWTS period. In the case of selenium, the reconfiguration from CWTS to the hybrid bioreactor-CWTS resulted in a statistically significant decrease in outflow concentration compared to that prior to the reconfiguration, with a lowest measured outflow concentration of 0.0019 mg/L and average of 0.004 mg/L (Figure 2; p < 0.001). This improvement in selenium treatment was expected based on the design targeting more favourable conditions for selenium reduction, such as lower DO and ORP resulting from the additional electron donors from microbial Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 554 Mine Closure 2015, Vancouver, Canada decomposition of the alfalfa and straw. In regards to copper, the lowest outflow concentration of the study was also measured during the hybrid bioreactor-CWTS phase (0.0075 mg/L); however, treatment effectiveness decreased briefly in all systems prior to this (Figure 2, 19 June data point). Even during this brief disruption, 86% of the total copper entering the system was removed from the water. This disruption in treatment effectiveness coincided with low DO and moderately reducing conditions in the water column being attained in all cells (DO < 0.5 mg/L, ORP < −80 mV, compared to > 3.5 mg/L and > +142 mV prior to adding alfalfa and straw). It should be noted that the leach test conducted on the alfalfa and straw prior to addition determined that there was no discernible release of copper from these organic materials. The cause of the disruption could not be determined definitively, with the disturbance being registered during a single sampling point and a maximum potential disturbance period of four weeks.  The tests conducted in this study demonstrate that the CWTS designed for Minto can be repurposed as a hybrid bioreactor-CWTS resulting in improved selenium and copper treatment within the same footprint (Figure 2), and without need for added construction. However, operating the system as a hybrid bioreactor-CWTS would require greater operational maintenance than it would in its original CWTS configuration. While the hybrid bioreactor-CWTS configuration improved selenium treatment and demonstrated the lowest outflow concentration of copper of the entire trial, the short upset of treatment performance during the transition is meaningful in terms of monitoring during site implementation. It should be emphasised that if a transition from a typical CWTS design to a hybrid bioreactor-CWTS style is performed, careful monitoring of the CWTS would be required, as some elements may have decreased treatment effectiveness for a brief period. During this timeframe, if concentrations of elements exceed regulatory guidelines, water could be recycled through the wetland, or a contingency water treatment plan could be put in place for the outflow over the period of treatment disturbance.  While per cent removal is often the most widely discussed aspect of system treatment performance, by itself it is not a valuable measure. There is a great difference between 99% removal with a starting value of 10,000 (leaving 100 units of a constituent), compared to 99% removal for a starting value of 10 (leaving only 0.1 units of a constituent). In fact, the per cent removal of a constituent in a system provides almost no information of value without the context of HRT and initial and final concentrations. As such, it is more appropriate to discuss and evaluate treatment performance in the context of reaction kinetics. The use of two cells in each series allowed for the opportunity to assess the amount of treatment occurring at different points in the pilot-scale CWTSs, and thereby for the calculation of removal rate coefficients to be developed for ranges of concentrations. As can be seen in Figure 3, most of the copper is removed in the first cell, with only a small Case Studies Mine Closure 2015, Vancouver, Canada 555 amount of treatment occurring in the second cell. Conversely, selenium treatment occurs in a more linear fashion, through both cells in the system. This information can be used to provide more accurate sizing and performance estimates of a full-scale CWTS.  Based on the treatment performance measured in this study, removal rate coefficients were calculated for the long-term closure scenario with low ammonia and nitrate concentrations. The removal rate coefficient of k = 0.01 for selenium can be applied between the concentrations of 0.002 and 0.012 mg/L,  k = 0.055 for copper between 0.036 and 0.150 mg/L, and k = 0.025 for copper between concentrations of 0.008 and 0.035 mg/L. At concentrations higher than 0.150 mg/L, a greater value of k is likely applicable. It remains to be experimentally determined through onsite demonstration-scale testing whether it is possible to achieve even lower outflow concentrations of selenium and copper with this wetland design. However, lower concentrations of both elements have been achieved with CWTS designed for other sites, suggesting it is thermodynamically possible (Haakensen et al., 2013). These removal rate coefficients can be used with confidence at a temperature of 20°C, and could be adjusted conservatively by using the Arrhenius equation. The Arrhenius equation is considered conservative in this circumstance because some of the rate-affecting processes for treatment in a CWTS are sorption, filtration and precipitation, which are not impacted by temperature in the same way as chemical and biological reactions. While plants are often thought of as a treatment pathway through uptake, this pathway is not sustainable for metals, as they can be released upon decomposition. However, plants can provide multiple benefits to the function of a CWTS, and uptake of elements can be minimised through appropriate designs that target mineralisation and decreased bioavailability. In these systems, copper treatment was highly correlated to sulphur deposition in soils, and not correlated to iron concentrations.  Mass balances were calculated for each pilot-scale system to assess the fate and distribution of elements removed from the water during the course of testing. For systems that were planted with both Carex and moss, 64–75% of the copper removed from the water was sequestered into the top six centimetres of sediment. In contrast, systems with only Carex planted and no moss had only 33% of the copper ending up in this top sediment fraction. For systems planted with both Carex and moss, 68–83% of treated selenium was sequestered to this top portion of the sediment, but less than 52% was in the sediment of the system planted with only Carex (actual percentage cannot be calculated, as the selenium was below the sediment detection level; the 52% assumes actual concentrations in the soil are equal to the detection limit). For both elements, only 1–3% of the total load removed from the water was found in the above water vegetation (Carex) at the end of pilot-scale testing. It was not possible to determine the concentrations of elements in moss, as the majority did not survive in the hybrid bioreactor-CWTS conditions. As such, the majority of copper and selenium treated by the system was sequestered to the top six centimetres of sediment, with only a small portion being taken up into the Carex. Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 556 Mine Closure 2015, Vancouver, Canada Although it was expected that moss would perform well in treating water through sorption, it was also expected to generate high DO concentrations in the water (through photosynthesis), which could negatively impact the targeted negative oxidation-reduction conditions necessary for sulphate and selenium reduction in the systems. Furthermore, if the CWTS was built without moss, it was expected to eventually include moss, as this was the natural state of wetlands found at Minto through the site assessments (reports 2013-0100-256 and 2013-0100-257 on the YESAB registry). Therefore, it was necessary to test whether it was beneficial or detrimental to have moss in the CWTS design.  All plants were harvested and transported from the Minto Mine with the above-ground biomass (leaves) removed before transplanting to ensure that measurements were reflective of actual uptake during this study. The Carex and moss both had a very high survival rate in the pilot-scale systems, with increase of stem count and total mass through the entire study (Figure 4). The plants were established for six weeks after transplantation with tap water running through the pilot-scale CWTSs before synthetic influent was started. The maximum average stem/leaf height was reached after approximately three months. During these first three months, the average stem/leaf height increased, while the total number of stems remained the same. It was only after these plants achieved the maximum height that they began to send up additional stems. The stem count was still increasing when the study ended, suggesting that maximum plant density had not yet been achieved and would likely take more than the length of a single growing season to establish (Figure 4). During system takedown, it was apparent that the Carex roots had spread throughout the cells, providing a strong foundation for the plants. Moreover, many of the roots were associated with sulphate-reducing zones, as was evident by black areas in the surrounding sediment.   It is important to the overall treatment system design to determine if plants may release elements during freezing and thawing of the water, as would occur through natural seasonal progression. A freeze-thaw trial was conducted in which Carex and moss were collected and sent for metals analysis both before and after being submerged in water from the wetland and freezing to −20°C followed by thawing to room temperature. It was found that Carex released 18–38% of accumulated selenium and 0–13% of accumulated copper based on the decrease in tissue concentrations after freezing and thawing. In contrast, moss was found to have more than a two-fold increase in the concentration of these elements after being frozen and thawed. It appears that through the freeze-thaw process, the moss has an enhanced sorption of these compounds, removing an additional amount from the water. Based on the pilot-scale system design and plant densities, Carex would theoretically release on average 0.14 mg of copper and 0.27 mg of selenium per square metre of wetland area, while moss would theoretically bind 511 and 11.8 mg/m2, respectively. This finding is important, as it demonstrates that despite some release of constituents by Carex through freeze-thaw, the moss will uptake an excess of what is released by the Carex, thereby assisting with treatment during the initial spring thaw before biogeochemical processes are at full capacity.  Case Studies Mine Closure 2015, Vancouver, Canada 557 Using genetic- and growth-based methods, bacterial communities during all CWTS pilot-scale testing phases were compared. The soil used to construct the pilot-scale systems had low bacterial diversity (Simpson’s reciprocal index), which increased as plants were established in the pilot systems (prior to influent running through the systems). Similarly, the relative abundance of SRB was very low in the initial soil and increased as plants were established in the systems. The high N and low N phases had similar bacterial diversity and an increase in relative abundance of SRB (genetic) and total heterotrophs (growth-based) compared to before influent was running through the systems. The hybrid bioreactor-CWTS period also showed an increase in bacterial diversity, and the highest relative abundance of SRB. The low N and hybrid bioreactor-CWTS phases had overall similar microbial populations, except that the hybrid bioreactor-CWTS had a higher relative abundance of SRB. This indicates that either: (1) the microbes present in the low N phases were well suited to the bioreactor-CWTS conditions, or (2) that the community was still in flux at the final sampling, and treatment may have continued to improve over time in the hybrid system as more anaerobic organisms flourished.  The results of the growth-based MPN analysis indicate an abundance of organisms capable of denitrification was present in the high N period. Although in lower abundance, denitrifying bacteria were present through the low N period and the hybrid bioreactor-CWTS period. The MPN for selenite-reducing bacteria ranged from hundreds of thousands to tens of millions of organisms per gram of soil and was similar between the different system designs and testing periods. The number of SRB steadily increased over time in all designs, starting with a MPN of approximately 1,000 organisms per gram of sediment, which is similar to that found associated with Carex through the site assessment (reports 2013-0100-256 and 2013-0100-257 on the YESAB registry), and increased by two orders of magnitude through the study. This increase is consistent with trends observed in the relative soil redox measurements of the pilot-scale systems becoming more negative over time, and is indicative of the system becoming more established and achieving the desired reducing conditions targeted for treating constituents of interest. Both genetic and growth-based methods found the presence and abundance of SRB were positively correlated with final sulphur concentrations in sediment. This shift in microbial communities (paired with the shift in population observed for denitrifying organisms) is indicative of the responsiveness and adaptability of the CWTSs to different influent chemistries. It has previously been shown that once established, the microbial communities in CWTSs are robust and maintain similar community compositions even after freezing and thawing (Haakensen et al., 2013). Several OTUs related to known SRB were found to correlate with sediment concentrations of copper (p < 0.001) and sulphur (p < 0.1), suggesting that their presence contributes to reduction of sulphate to sulphide, which then binds to copper and precipitates into the sediment. This is supported by the appearance of black areas surrounding the Carex roots and black precipitates on the sediment surface observed during system takedown. This copper-sulphide mineralisation mechanism is ideal for treatment, as it produces a stable form of copper in the sediment. We can tell from the genetic sequencing that these same types of SRB were found in samples collected during the August 2013 Minto site visit, particularly associated with Carex roots. It is likely that these beneficial microbes were brought to the pilot-scale system along with the Carex during transplanting. This indicates that despite differences in sediment, water and environmental conditions between the harvesting location and the pilot-scale CWTS, these microbes are robust and can be further encouraged to thrive and perform beneficial functions by being provided the ecological niches that they prefer in the context of the CWTS (e.g. Carex roots, low ORP, and low DO).  Based on data collected during pilot-phase testing, we conclude that the combination of Carex and moss is the ideal CWTS design for long-term, sustainable copper and selenium treatment for the Minto Mine. In this design, Carex provides stability to the wetland soil and sediment with extensive root structure, additional treatment capacity based on the roots drawing water down into the sediment, sulphate-reducing zones and renewal of organic carbon in the sediment. The moss provides sorption and uptake of elements even through freeze-thaw events. Moreover, both Carex and moss are known to be accreting, peat-forming plants, which Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 558 Mine Closure 2015, Vancouver, Canada is beneficial to soil quality in the CWTS over time. Using two cells in each series permitted removal rate coefficients to be developed for ranges of concentrations, allowing for more accurate sizing and performance estimates of a full-scale CWTS. Phase 4 of the design and optimisation program is underway, with a demonstration-scale CWTS built onsite at the Minto mine in the fall of 2014. Based on research findings presented in this paper, the demonstration-scale CWTS is constructed as a free water surface flow CWTS planted with both Carex and moss. Biochar was not incorporated in the soil, as the findings of the pilot-scale studies showed that it did not result in improved treatment or greater abundance of beneficial microbes.  We would like to acknowledge Jenny Liang, Qingxiang (Mark) Yan and Emily Charry Tissier, all of Contango Strategies, for work performed during pilot-scale testing.  Blodgett, R. (2010) US Food and Drug Administration, Bacteriological analytical manual, Appendix 2: Most probable number from serial dilutions. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., Fierer, N., Peña, A.G., Goodrich, J.K., Gordon, J.I., Huttley, G.A., Kelley, S.T., Knights, D., Koenig, J.E., Ley, R.E., Lozupone, C.A., McDonald, D., Muegge, B.D., Pirrung, M., Reeder, J., Sevinsky, J.R., Turnbaugh, P.J., Walters, W.A., Widmann, J., Yatsunenko, T., Zaneveld, J., and Knight, R. (2010) QIIME allows analysis of high-throughput community sequencing data, Nature Methods, Vol. 7, pp. 335–336. DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K. Huber, T., Dalevi, D., Hu, P., and  Andersen, G.L. (2006) Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB, Applied and Environmental Microbiology, Vol. 72, pp. 5069–5072. Edgar, R.C. (2010) Search and clustering orders of magnitude faster than BLAST, Bioinformatics, Vol. 26, pp. 2460–2461. Faulkner, S.P., Patrick, W.H. Jr. and Gambrell, R.P. (1989) Field techniques for measuring wetland soil parameters, Soil Science Society of America Journal, Vol. 53, pp. 883–890. Haakensen, M., Terry, M.W., Pittet, V., Castle, J.W. and Rodgers, J.H. Jr. (2013) Effects of freeze-thaw and biochar on sequestration and localization of elements within oxidizing and reducing pilot constructed wetland treatment systems, SETAC North America 34th Annual Meeting, 17–21 November 2013, Nashville, USA, Presentation 593, recording available online at Haakensen, M., Pittet, V., Spacil, M.M., Castle, J.W. and Rodgers, J.H. Jr. (2015) Key aspects for successful design and implementation of passive water treatment systems, Journal of Environmental Solutions for Oil, Gas, and Mining, Vol. 1, pp. 59–81. Haines, W.P. and Renwick, J.A.A. (2009) Bryophytes as food: Comparative consumption and utilization of mosses by a generalist insect herbivore, Entomologia Experimentalis et Applicata, Vol. 133, pp. 296–306. Hawkins, W.B., Rodgers, J.H. Jr., Gillespie, W.B. Jr., Dunn, A.W., Dorn, P.B. and Cano, M.L. (1997) Design and construction of wetlands for aqueous transfers and transformations of selected metals, Ecotoxicology and Environmental Safety, Vol. 36, pp. 238–248. Huddleston, G.M. III, Rodgers, J.H. Jr. (2008) Design of a constructed wetland for treatment of copper-contaminated wastewater, Environmental Geosciences, Vol. 15, pp. 9-–9. Kadlec, R.H. and Wallace, S.D. (2009) Treatment wetlands, 2nd edition, CRC Press Taylor & Francis Group, Boca Raton, USA. Klindworth, A., Pruesse, E., Schweer, T., Peplies, J., Quast, C., Horn, M. and Glöckner, F.O. (2013) Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies, Nucleic Acids Research, Vol. 41, pp. e1. Lefrançois, E., Quoreshi, A., Khasa, D., Fung, M., Whyte, L.G., Roy, S. and Greer, C.W. (2010) Field performance of alder-Frankia symbionts for the reclamation of oil sands sites, Applied Soil Ecology, Vol. 46, pp. 183–191. Longton, R.E. (1997) The role of bryophytes and lichens in polar ecosystems, in Ecology of arctic environments: 13th special symposium of the British Ecological Society, 1st Edition, Blackwell Science Ltd., Oxford. Lozupone, C., Lladser, M.E., Knights, D., Stombaugh, J. and Knight, R. (2011) UniFrac: An effective distance metric for microbial community comparison, ISME Journal, Vol. 5, pp. 169–172. McDonald, D., Price, M.N., Goodrich, J., Nawrocki, E.P., DeSantis, T.Z., Probst, A., Andersen, G.L., Knight, R., and Hugenholtz, P. (2012) An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea, ISME Journal, Vol. 6, pp. 610–618. McMurdie, P.J. and Holmes, S. (2013) phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data, PLoS ONE, Vol. 8, p. e61217. Murray-Gulde, C.L., Bridges, W.C. and Rodgers, J.H. Jr. (2008) Evaluating performance of a constructed wetland treatment system designed to decrease bioavailable copper in a waste stream, Environmental Geosciences, Vol. 15, pp. 21–38. Rodgers, J.H. Jr. and Castle, J.W. (2008) Constructed wetland systems for efficient and effective treatment of contaminated waters for reuse, Environmental Geosciences, Vol. 15, pp. 1–8. Case Studies Mine Closure 2015, Vancouver, Canada 559 Siddique, T., Zhang, Y., Okeke, B.C. and Frankenberger, W.T. Jr. (2006) Characterization of sediment bacteria involved in selenium reduction, Bioresource Technology, Vol. 97, pp. 1041–1049. Suren, A.M. and Winterbourn, M.J. (1991) Consumption of aquatic bryophytes by alpine stream invertebrates in New Zealand, New Zealand Journal of Marine and Freshwater Research, Vol. 25, pp. 331–343. Taylor, C.R. (2009) Selecting plant species for optimize wastewater treatment in constructed wetlands, MSc Thesis, Montana State University, Bozeman, USA.   Process-driven design and piloting of a site-specific constructed wetland for copper and selenium treatment in the Yukon M. Haakensen et al. 560 Mine Closure 2015, Vancouver, Canada  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items