Biogeochemistry and biologic sulphate reduction in Hammaslahti pit lakes, eastern Finland

Marja Liisa Räisänen¹, Jarno Mäkinen² and Malin Bomberg², ¹Geological Survey of Finland, P.O.Box 1237, FI-7021 Finland, marja-liisa.raisanen(at)gtk.fi; ²Technical Research Center, P.O.Box 100, FI-02044 VTT Finland

Introduction of the mine site

The closed Hammaslahti copper mine is located in Joensuu (former Pyhäselkä municipality), about 25 km south of the Joensuu city center, Eastern Finland. The ore was discovered in 1966. Open pit mining was started in the southern (S) orebody in the winter of 1973, and was followed by underground mining from 1976 to 1986 (Pelkonen et al. 1973, Karppanen 1986). The open pits were named S, N (northern) and Z pits (Fig. 1). An ore body with exceptional Zinc concentration was mine from the Z pit from 1983 and 1984. At present, the mine area includes a tailings facility, a waste rock heap and three open pits filled with water (Fig. 1). The former settling pond was dried for use as a shooting range by the Pyhäselkä hunting club. The mine buildings (concentrator, office) were repaired for use by a private company (saw mill).

Figure 1. Location of the study site in the closed Hammaslahti mine area, eastern Finland. White arrows mark the water flow direction from the Z pit and wetland complex (see also Fig.2), orange arrows seepage water (acid) directions in the area surrounding the tailings facility, blue arrows the water flow direction of the receiving Iiksenjoki River. Airphoto © National Land Survey of Finland.

Geology, ore mineralogy and chemistry

The Hammaslahti copper ore was hosted by Lower Proterozoic metaturbidites (arenites, argillites) of the Höytiäinen Province (Karppanen 1986, Loukola-Ruskeeniemi et al. 1992). The metaturbidites hosting the deposit have been subjected to intense silicification and chloritization (Loukola-Ruskeeniemi et al. 1992). In deeper parts of the mine, the black schist is host to remobilized sulphides (Karppanen 1986). The country rocks are quartz-chlorite-hornblende schist (volcanic origin), quartz-chlorite schist, chlorite schist, tremolite-carbonate rock, feldspathic greywacke and conglomerate, phyllite and black schist (Loukola-Ruskeeniemi et al. 1992).

The S (southernmost pit in Fig. 1) orebody accounted for 70% of the total ore reserves (Karppanen 1986). The major sulphides were pyrrhotite and chalcopyrite with minor sphalerite (Pelkonen at al. 1973). The upper part of the orebody was slightly banded massive sulphide ore, which gradually changed with depth to a remobilized network of veins and disseminations (Karppanen 1986, Loukola-Ruskeeniemi et al. 1992). A breccia ore shooting from the S orebody into the hanging-wall black schist consisted of chalcopyrite, pyrrhotite and pyrite (Loukola-Ruskeeniemi et al. 1992).

The smaller N (northern, central pit in Fig. 1) orebody below the S orebody consisted of chalcopyrite. Locally there were very high grade but low volume concentrations of chalcopyrite, and low concentrations of sphalerite. (Loukola-Ruskeeniemi et al. 1992)

The Z (zinc) northernmost orebody comprised the pyrite (western limb) and sphalerite ores (eastern limb). The western ore limb consisted of pyrite and pyrrhotite together with some chalcopyrite and sphalerite. The edges of the massive sulphide lens were rimmed by secondary pyrrhotite and concentrations of remobilized chalcopyrite. The eastern ore limb consisted of pyrrhotite, sphalerite and chalcopyrite that form a banded breccia ore zone. The breccia ore had minor pyrite as porphyroblasts. In places, quartzite bands occured in turns together with pyrrhotite and galena disseminations. In folded and sheared sites, pure quartz veins alternated with sphalerite, chalcopyrite, pyrrhotite and galena bands. (Hämäläinen 1987)

The ore deposits contained 1.11 wt-% Cu and 1.26 wt-% Zn on average (Puustinen 2003). According to Karppanen (1986), the gold grade was 0.5-2.0 g/t (with a maximum 10 g/t) in the S and Z orebodies of higher Zn grade, whereas in the chalcopyrite ore, the gold content was 0.1-0.2 g/t on average (Loukola-Ruskeeniemi et al. 1992).

Mining and milling processes

A total 7.9 Mt of rocks and about 5.6 Mt of ore were mined from the underground mine and three open pits (Puustinen 2003). The annual production of ore was about 0.4 Mt.

The flotation plant mainly produced copper concentrates from 1973-1982, and also zinc concentrates from 1983-1986 (Hämäläinen 1987). The gold was recovered in copper concentrate (Karppanen 1986). The pH of the flotation slurry (copper ore) was modified with caustic lime to pH 11.5 (Kennedy 1985). Sodium amyl xanthate was used as a collector and Montanol (pine oil) as a frother (Pelkonen et al. 1973, Kennedy 1985). There is no published data on the zinc ore flotation method used at the Hammaslahti mill. Obviously, copper sulphide tailings were fed to the sphalerite flotation circuit, which included the separation of Fe sulphides.

Tailings from the flotation process were pumped to a thickener, the overflow from which was returned to the concentrator’s process water reservoir. The underflow (tailings slurry) was pumped to the tailings impoundment, from which clarified water was recirculated via the pump station to the process water reservoir. The mine conducted complete recycling of process water. Only during very dry seasons was additional water used from other sources, such as Myllylampi Lake (east of the map in Fig.1). In the case of exceptionally abundant rainfall in the spring, some water from the process was discharged into surrounding watercourses (Iiksenjoki River).

Mine waste management

Tailings characterization

The amount of tailings is about 5.3 Mt. Tailings consist mainly of quartz, chlorites, tremolite, plagioclase and sulphides, especially Fe sulphides, pyrrhotite and pyrite. According to GTK’s drilling data analysed from 2000-2004, the tailings consist of 3.76 wt-% S, 760 mg/kg Cu, 1,230 mg/kg Zn, 55 mg/kg Co and about 30 mg/kg Cr and Pb on average (Table 1, Tenhola & Räisänen 2006). Based on the pretty high total sulphur content (>1%) and low primary carbonate mineral content, it can be assumed that the tailings are acid generating mine waste. This is confirmed by the acidification of drainage waters by tailings seepage (Räisänen et al. 2003).

Table 1. Mean, median, minimum (Min) and maximum (Max) concentrations of S and trace metals characteristics for the Hammaslahti tailings (see also Tenhola & Räisänen 2006). The sulphur content was measured with a pyrolytic method (Leco) and concentrations of the chalcophile metals with the ICP-OES method using hot Aqua Regia for the sample digestion. Altogether 183 samples were analysed.

Waste rocks characterization

The amount of waste rocks is about 2 Mt. Waste rocks consist of metagreywacke, phyllite, quartzite, mica and quartz-chlorite schists, and tremolite bearing black schists (See Hämäläinen 1987, Loukola-Ruskeeniemi et al. 1992). Several waste rocks contain sulphide minerals (mainly Fe sulphides) and therefore they can be classified as acid generating rocks. Earlier findings of acidic drainage waters discharged from the waste rock heap confirm this interpretation (Räisänen et al. 2003). There was no available chemical data on waste rocks.

Design and basal structures of the waste facilities

The main mine waste area is a tailings impoundment of 30 hectares. The facility is constructed on a bog, north-east of the open pits (Fig. 2). The initial starter dam was constructed with sandy till, raised later with coarse tailings, and covered with waste rock boulders on both the wet and dry banks. The impoundment is divided into two ponds with a mid-dam (See Fig. 2a, Tenhola & Räisänen 2006). The water clarified in the eastern impoundment segment was pumped for final clarification into a settling pool of 4 hectares, south-east of the impoundment (Fig. 2). The pool water was mainly re-used in the process at the concentrating mill. Floodwaters in spring (snow and ice melting), and during heavy rains in summer and autumn were discharged via an open ditch, east of the facility into the Iiksenjoki River. Overall, the need for extra fresh water pumped from the Myllylampi Lake, east of the mine area was minor in quantity. (Pelkonen et al. 1973)

The pad structure of the tailings consists of a peat layer, 0.2-0.8 m thick which has become compressed under the tailings (Räisänen et al. 2003). According to the drilling observations, the peat is underlain by glaciolacustrine silt and glacial till, respectively. The chemical data of the drilling cores showed that the peat effectively traps contaminants dissolved from the tailings (Tenhola & Räisänen 2006). The waterproof, compact basal structure results in a hydraulic pressure in the tailings that gradients towards the dams in the north, west and east (Räisänen 2003, Tenhola & Räisänen 2006). In the south, the tailings lie on rocky terrain covering shallow till. Due to the compact basal structure, groundwater under the facility stays uncontaminated, but the seepage contaminates surface waters and groundwater in the bog surrounding the facility (Räisänen et al. 2003).

The settling pool was dammed on the bog bordering the tailings impoundment in the southeast (Fig. 2a). In contrast to the tailings facility, peat from the bottom of the pool was removed down to till. In addition, a layer of sulphide bearing waste rock aggregate was added on top of the till. After the mine closure, the settling pool was drained for use as a shooting range (Fig. 2b).

The main waste rock heap is about 5 hectares and is located west of the S open pit. There is also a smaller heap on the southern slope of the S pit (Figs. 1 and 2). Waste rocks in the eastern heap lie partly on rocky terrain and in the west and north are on shallow till. The small heap is partly backfilling the southern corner of the S pit.

Figure 2. Facilities of the Hammalahti mine (a) in 1974 and (b) after closure in 1996, eastern Finland. AIrphotos @ National Land Survey of Finland.

State of the mine closure

Open pits and water management

The underground mine was gradually filled by groundwater after finishing the quarrying. At the beginning of the 1990’s the three open pits (S, N, Z) were allowed to fill with subsurface water and rainwater. Waste rocks were backfilled into the southern end of the S pit and in the bottom of the N pit. Presumably, the pits are joined by underground tunnels, and the waters are discharging from the northernmost pit, Z pit. Originally, the waters overflowed into a ditch bordering the cultivated field, northwest of the Z pit (Fig. 2b). In 2002, the discharge was turned to flow out from the southeast corner of the Z pit into a complex of wetland pools (Fig. 3). The overflow site was filled with carbonate rock boulders (containing dolomite and calcite minerals, and with diameters from 5-15 cm). The receiving pool was originally a till excavation, where shallow water near the pit edge was lined with peat to promote aquatic plant growing. Subsequently, the water flows into a broad ditch lined with peat, and seeps through a road bank covered with cobble to boulder size carbonate rocks into a two-segment wetland pool. The pool was dammed with till on a mire. The wetland pool also receives seepage from the south western part of the tailings heap. The section of the pool receiving the seepage was covered with carbonate rock (diameter of 5-15 cm) overlying the peat. Finally, waters from the wetland pool flow via a pipe into an open ditch circling in the bog and into the Iiksenjoki River (Fig. 1).

Figure 3. Locations of the Z open pit and wetland pool complex in the closed Hammaslahti mine area, eastern Finland. White arrows mark the direction of water flow starting from the pit through the wetland complex and ditch network in the bog. Orange arrows mark the direction of seepage water flow in the western side of the tailings impoundment. Airphoto © National Land Survey of Finland (See also Fig. 1).

The aftercare treatment of pit waters consisted of reducing sulphate and metals with the addition of a bacteria matrix. First in 1998, and again in 2000, a total of 300 m³ of pig manure sludge was added to the water of the N pit, followed by the addition of 150 m³ of wood park chips in 2002 (Vestola & Mroueh 2008). In 2004, more pig manure sludge was added, followed by approximately 23,000 litres of ethanol in 2005. It was assumed that the slow decay of the chips will release carbon as an energy source for bacteria.

At the end of the 1990’s and the beginning of 2000, pH of the overflow water from the Z pit was less than four in summers and sometimes into later autumn, but over six in the winter (ice covered) and May (Table 2, see also Räisänen et al. 2003). During the same period, pH of the surface water in the S pit was also acid (pH<4), but pH of the deep water was over 6 (Table 2). The pH of the water in pit N was mostly over six year-round in 2000 and 2001, with the exception of water at 25 m depth, and once at 37 m depth, which was less than 6 (5.4-5.9).

Table 2. pH variation of the pit waters with depth in March and August 2000 and in March 2001, Hammaslahti. The pH was measured in the laboratory of Savo-Karjalan ympäristötutkimus Oy (Unpublished monitoring data of Outokumpy Oyj).

Mine waste heaps and water management

The main decommissioning actions of the tailings facility were drying of the topmost tailings, spreading till cover, and vegetating. The water table in the tailings impoundment was gradually decreased by pumping free water into the settling pool. After treatment with lime, the pond water was discharged via a ditch network into the Iiksenjoki River (Fig. 1). In 1996, the settling pond was drained for shooting range construction (moose shooting range, clay-pigeon shooting).

The tailings facility was covered in three phases. At the beginning of 1990’s the major part of the tailings facility was covered with a till layer from 10 to 60 cm thick. The western, uncovered part of the tailings was used as an airstrip for light aircrafts. Due to dust, the airstrip was removed from use, covered with till, and vegetated with grass in 2000. In 2002, peat and carbonate-rock chips were spread on the till. In 2004, the whole tailings area was fertilized with pig manure sludge, which led to denser vegetation, and markedly increased tree growth (pine, birch). (See also Räisänen et al. 2003, Tenhola & Räisänen 2006)

Originally, seepage waters from the tailings facility flowed via a ditch network from the bog into the Iiksenjoki River (Räisänen et al. 2003). In 2003, a sand road bound by a broad ditch was constructed to around the tailings facility (Fig.1). The ditch efficiently collects tailings seepage waters and provides gentle steps down to a pool with two segments. The pool water flows through a pipe under the road into a ditch lined with carbonate rocks, and then into the Iiksenjoki River. The water from the Z pit flows into a wetland complex (Fig.2). Water discharges from the wetland complex via a ditch surrounding the bog, into the same segmented pool, and finally into the Iiksenjoki River (Fig.1).

The water data from 2000-2001 and 2012 show that the pH of the drainage waters stayed <3.5, on average (Table 3). In addition, the drainage waters had pretty high sulphate, Fe, Al and Mn concentrations (1,230 mg/l, 86 mg/l, 9.6 mg/l and 7.2 mg/l, respectively). In 2012 concentrations of sulphate and Fe remained high despite changes in the drainage of the bog surrounding the tailings facility (Viitasalo 2013). On the contrary, concentrations of Al and trace metals typical for the seepage decreased. Concentrations of sulphate and Mn of the Iiksenjoki River downstream of mine site discharge remained unchanged between 2000-2001 and 2012 data, whereas the downstream Fe concentration markedly increased in the same period. A similar increasing trend was observed for Al, but not for Zn. The Zn content in the drainage waters was about 10-fold higher in 2000-2001 (3.6 mg/l) than in 2012 (0.2 mg/l), while in the same period the Zn concentration in the upper Iiksenjoki River decreased by two orders of magnitude. Concentrations of Cu, Ni, Co, Pb and Cd, which were pretty low during both sampling periods, were ≤ 0.1 mg/l in discharge, and even lower in the Iiksenjoki River.

Table 3. Physical and chemical composition of the drainage waters and Iiksenjoki River waters, the closed Hammaslahti mine, eastern Finland. The drainage waters were sampled from a ditch collecting waters from the drain network that surrounds the tailings facility and the northernmost open pit (Z), which discharge into the Iiksenjoki River (Fig. 1).

The waste rock heaps stayed uncovered after mine closure. At present, banks of the main waste rocks heap are growing leaf trees (See Fig. 1). After mine closure, a small portion of the waste rock was backfilled into the N and S pits, and locally used for road and pier constructions. At the beginning of 2000, the re-use was stopped due to increased understanding of acid generating risk by the regional environmental authority. Originally, seepage from the main heap in north flowed into a ditch that steps down to the till excavation, south of the Z pit. In 2004, the seepage was drained to flow into the N pit. In central and eastern parts of the heap, the seepage flows into the S pit. According to GTK data from 2000 and 2001, the seepage from the heap into the ditch was acidic, with pH less than 4 (Räisänen et al. 2003).

Description of study sites and targets

The Hammaslahti case study introduces passive water treatment technologies applied in two open pits and pit overflow waters. The study sites include the open pits N and Z, which filled with water after mine closure, and the partially constructed wetland pool complex, into which the waters overflow from the south western corner of the northernmost Z pit (Fig. 1). It is expected that the water from N pit flows via an underground tunnel into the Z pit.

The main targets of this case study are to examine the physical, chemical and biological quality of the pit waters, remediation capacity based on sulphate reduction and the subsequent precipitation of metal sulphides. Furthermore, it is of interest to figure out the capacity of the wetland pool complex as a final treatment of overflow water before its discharge into a natural waterway.

Materials and research methods

Materials for the N and Z pit study consist of physical measurements of whole pit water columns (measurements taken at depth intervals of every 1 to 2 meters), physical, microbiological and chemical analysis of water samples from selected levels, and mineralogical and chemical analysis of bottom sediments. Materials for the wetland pool complex study consist of physical measurements from selected pool water sites, and the chemical analyses of water samples from the main wetland segments and outflow water sites. Figure 4 shows the sampling sites and sites for physical water quality measurements.

Figure 4. Water sampling sites, the Hammaslahti closed mine area, eastern Finland. Bottom sediment samples were taken close to the pit water sampling sites in the Z and N pits. Airphoto © National Land Survey of Finland.

Water sampling and sample pre-treatment

The pit water samples were taken twice a year in 2014, i.e. in March 23^th (N pit) and 24^th (Z pit) and on September 8^th (N pit) and 9^th (Z pit). At sampling sites, the water depths in the N pit varied between 38 and 41 m and in the Z pit between 45 and 46 m. The samples were collected with a Limnos water sampler separately from three water depths (1 m, 20/22 m and 37/40 m) at one N pit site, and from four water depths (2 m, 30 m, 35 m and 45 m) at one Z pit site (see sites in Fig. 4). Duplicate water samples were only taken from the depth of 45 m in the Z pit. From each Limnos water batch two subsamples were collected (one litre and one half litre) in PEH, polyethylene 1.5 litre bottles.

The water quality of the overflow from the Z pit, downstream in the wetland complex, and further in the receiving bog ditch, was measured only on September 9^th 2014. The water samples were collected from the Z pit overflow site, both constructed wetlands (wetland 1, wetland 2), and two receiving ditch sites only in the autumn sampling round (Fig. 4). Volumes of two plastic bottles (0.5 l, 1 l) fixed onto a pole were taken at surface water sampling sites (Fig. 4).

Half a litre sample bottles taken from the pit waters and surface waters were used for determinations of anions and solid content. Subsamples from one litre volume samples were filtered on-site through a GD/XP 0.45 µm-filter into HDPE bottles as follows:

• 50 ml subsamples for the analyses of 34 dissolved elements. After filtering, the subsamples were acidified on-site by adding 0.25 ml supra-pure nitric acid.
• 100 ml subsamples for analysis of dissolved Fe²⁺. After filtering, the subsamples were acidified on-site with 4 ml of hydrogen acid.
• 100 ml subsample for analysis of dissolved organic carbon content (DOC). After filtering into PEH bottles the subsamples were acidified with 1 ml concentrated phosphorus acid.

In addition, unfiltered 100 ml subsamples taken in PEH bottles were separated from the 1-l samples and acidified on-site with 1 ml concentrated phosphorus acid. These were used for the total organic carbon content.

Physical measurements of water samples

Alkalinity was measured on-site from the remainder of the 1-l samples using a Hach digital titrator with 0.1600 N or 1.600 N H₂SO₄ to an end point of 4.5.

Temperature, pH, redox potential, electrical conductivity (EC), oxygen concentration and oxygen saturation of waters at each sampling site were measured on-site with a multiparameter field meter (YSI 600 multiprobe system). Pit waterbeds were measured starting from surface layer, at a depth of 1 m in the N pit and at 2 m in the Z pit, and continued downwards at intervals of 1 or 2 meters.

Pit sediment sampling method

Sediments from the bottom of the N and Z pit were collected by disturbing fine particles from the bottom of the pit into the water with the Limnos sampler and then filling the Limnos cylinder with the water-sediment suspension. Several suspension batches per pit were collected into a plastic bucket and solids were let to settle overnight. During settling, the buckets were covered by black plastic bags to avoid oxidation. After settling, clarified water was decanted from the bucket and the remaining sediment suspension was decanted into a 2 litre glass bottle. Glass bottles were sealed with corks and placed into a cool box for mailing to the VTT laboratory in Espoo.

Chemical and mineralogical analysis

Waters

Concentrations of anions (Br, F, Cl, NO₃, SO₄) were determined with ion chromatography and concentrations of 34 elements were analyzed using ICP-OES or MS-ICP. Concentrations of dissolved organic carbon and total carbon were measured with a C analyzer. The content of solids was determined by gravimetric method (SFS-EN 872:2005). Concentrations of Fe²⁺ were measured with a spectrophotometer. All the laboratory analyses were carried out at the FINAS-accredited testing laboratory of Labtium Ltd.

Sediments

Sediment samples were freeze-dried. Total element concentrations were determined with the semi-quantitative XRF method (by VTT Expert Services Ltd). Easily leachable and exchangeable element concentrations were analysed with a 1 M NH₄ acetate, pH 4.5 extraction method using a 1:10 solid to solution ratio (Räisänen et al. 1997). Acetate extractable element concentrations were determined with ICP-OES (by the FINAS-accredited testing laboratory of Labtium Ltd). The freeze-dried sediment samples were analysed for crystalline compounds with XRD and for organic matter with FTIR and TGA (by VTT Expert Services Ltd).

Microbiological analysis

For the characterization of microbial communities in the pit waters and sediments, microbial DNA was isolated from the samples using commercial DNA extraction kits. The total concentration of bacteria in the pit water was estimated by a DNA based qPCR method, where the number of bacterial taxonomical marker genes for the ribosomal small subunit (ssu) was calculated (as described in Tsitko et al. 2014). The bacterial community composition was determined by characterizing the whole community profile of the bacterial ribosomal ssu genes using high throughput amplicon sequencing on the Iontorrent platform. The sequence data was alalyzed with the QIIME pipeline (Caporaso et al. 2010) using the Greengenes database (DeSantis et al. 2006) for identification of sequence reads. The microbiology was assessed from the samples collected in March 2014.

Results

Physical quality of waters in the N and Z pits

Waters of the N pit were markedly more acidic (pH<6) than those of the Z pit (pH>6, see Figs. 5 a-b and Table 4). In both pits, the deeper waters were reducing (redox<0 mV), whereas surface waters were more or less oxidizing (Fig. 5). In the spring, when the N pit was covered with ice, acid generation was more notable at depths from 20 m to 22 m than above (Fig. 5a). In September when the pit was open, the acid generating layer (pH 4) thinned close to the depth around 20 m (See Table 4). In contrast to the N pit, the pH of waters in the Z pit did not markedly vary according to depth (Fig. 5b). However, in March the pH decreased slightly below the depth of 30-31 m, relative to surface in mid waters, while rising slightly with depth in September. In contrast to the pH of surface water in March, surface water pH in September was high, at 7.2, decreased with depth to near 6.3 in mid waters, and slightly increased (6.4-6.5) with depth below 32 m. In the Z pit, the redox potential decreased below 0 mV in deeper layers, i.e. below 35 m in the both seasons.

Dissolved oxygen from surface to deeper layers of both pits was directly related to redox potential (Fig. 6, Table 4). Reducing water layers lacked oxygen. In March, anoxic water occurred below the depth of 23-24 m in the N pit and below the depth of 31-32 m in the Z pit. In September, the deeper waters of both pits had, however, some dissolved oxygen ranging from 0.1 mg/l to 0.4 mg/l. It is worthwhile to notice that in the N pit, the oxygen concentration rapidly decreased from the surface down to the depth of 5 m and then stayed rather constant down to 21 m, below which was a final decrease (Fig. 6a). In contrast, the oxygen concentrations of waters in the Z pit stayed constant from the surface down to the depth of 32 m in March, whereas in September, the oxygen concentration rapidly dropped from surface water down to the depth of 5 m and then stayed more or less constant before a final decrease below the depth of 31-32 m (Fig. 6b).

Figure 5. Distributions of pH and redox in waters of (a) the N and (b) Z pit in spring (March) and autumn (September) in 2014, the Hammaslahti closed mine area, eastern Finland.

Figure 6. Distribution of oxygen concentrations (mg/l) in waters of the (a) N and (b) Z pit, the Hammaslahti closed mine area, eastern Finland.

The electrical conductivity of the waters was on average lower in the N than Z pit (Fig. 7, Table 4). In both pits, the deeper waters had conductivities twice as high as the surface and middle depth waters, following the decrease in redox potential rather than the change in pH with depth. Furthermore, in March the conductivities were markedly higher than in September in both pits.

In the N pit, waters had hardly any alkalinity due to acidity in surface and mid waters in March (Table 4). Both in March and September, deeper waters (at depth of 37-40 m) had some alkalinity (0.2-0.3 mmol/l). In contrast, the alkalinity of waters in the Z pit varied largely, between 0.5 and 5.6 mmol/l. In both pits alkalinity increased, pH rose and redox potential became reducing (below<0 mV) with depth.

The amount of solids was very low in the surface and mid water layers at both pits (Table 5). The greatest concentration of solid particles was measured from deeper water, and was about 130 mg/l in the N pit, and 30-120 mg/l in the Z pit. In the Z pit, the concentration of solids was much greater in September than in March, whereas in the deeper waters of the N pit the occurrence of solids was greatest in March. Furthermore, the increase in solids followed the decrease in redox potentials. The duplicate sample from the deepest water in the Z pit had almost ten times more solid particles than the primary sample. This is assumed to be caused by the mixing of fine particles (through suction) from bottom sediments into the water when lifting the sampler upwards.

Fiure 7. Distribution of electrical conductivities (mS/m) in waters of the (a) N and (b) Z pit, the Hammaslahti closed mine area, eastern Finland.

Table 4. Physical properties of the waters at the depths sampled in the N and Z pits, the closed Hammaslahti mine area, eastern Finland.

Chemical water quality of the waters in the N and Z pits

The chemical quality of water in the N pit is estimated on the basis of water chemistry at three depths, surface water (1 m), mid water (22 m in March and 20 m in September) and bottom water (40 m in March and 37 m in September). The selected water depths represent major changes in pH and redox potentials that presumably indicate microbial sulphide oxidation or iron and sulphate reduction (See Figs. 5-7).

The total concentration of organic carbon varied between 1.4 and 12 mg/l in waters of the N pit and between 1.6 and 5.6 mg/l in those of the Z pit (Table 5). In contrast to the Z pit, the amount of organic carbon was almost ten times greater (8.6-12 mg/l) in the surface water than in the mid and deeper waters (1.4-3.7 mg/l) of the N pit. In the Z pit, organic carbon concentrations in the surface and mid waters were only one third of those in deeper waters. The overall, majority of organic carbon was in soluble form.

The major anion in the pit waters was sulphate. Sulphate concentrations of the waters were on average greater in the N than Z pit. Furthermore, in both pits sulphate concentrations increased with depth, being highest in deeper waters. Comparison of sulphate concentrations (S-SO₄) with the sulphur concentrations measured by ICP-OES shows that most soluble sulphur was in sulphate form in both pits. However, some samples, especially in March, had less S-SO₄ measured by IC than S measured by ICP-OES. This feature was more characteristic for deeper waters of the Z than the N pit. Furthermore, the difference between the concentrations was greater in reduced (negative redox) than oxidized waters and presumably indicates the presence of reduced sulphur (S^2-), especially in deeper waters. The interpretation must include consideration for the difference in the sample preservation. Samples collected for ICP-OES determination of S were preserved with nitric acid. Samples collected for the determination of SO₄ were not preserved before IC analysis and were therefore more susceptible to chemical transformation (e.g. oxidation, adsorption, precipitation) than the filtered and acidified samples.

The concentrations of chloride ranged from 2.2 mg/l to 3.3 mg/l in the waters of the N pit and from 2.4 mg/l to 10.2 mg/l in those of the Z pit (Table 5). Generally, the amount of Cl was somewhat greater in March than in September, which seems to follow the decrease in pH in March observed in both pit waters. An exception was observed in one of two samples collected from the depth of 45 m in the Z pit. The primary sample had markedly less Cl in March (2.7 mg/l) than in September (9.5 mg/l) compared to the duplicate sample (10.2 mg/l and 9.1 mg/l, respectively).

Concentrations of F and NO₃ were low and mostly below the lowest detection limits (F<1 mg/l, NO3<2 mg/l, Table 5). Overall, the small concentrations that were recorded for samples which had less interference during the IC measurements are considered to be more or less representative.

Table 5. Concentrations of organic carbon and anions in water samples taken from the N and Z pit, the closed Hammaslahti mine area, eastern Finland. Key: < below the lowest detection limit.

Concentrations of alkaline earth (Ca, Mg) and alkaline (Na, K) metals were on average greater in waters of the Z (Ca 394 mg/l, Mg 93 mg/l, Na 19.7 mg/l, K 17.2 mg/l) than N pit (Ca 49.2 mg/l, Mg 19.3 mg/l, Na 3.19 mg/l, K 6.28 mg/l). In both pits, the concentrations almost doubled from surface to deeper waters and they were slightly greater in March than in September (Table 6). The exception to the seasonal variation was the reversed distribution of Ca and Mg in the surface water of the N pit and in the deepest waters of the Z pit (primary samples, not duplicates).

Compared to the alkaline earth and alkaline metals, aluminium was distributed vice versa (Table 6). On average, waters in the N pit had more soluble Al (2.17 mg/l in March, 1.70 mg/l in September) than those in Z pit (0.004 mg/l and 0.17 mg/l, respectively). The greatest Al concentrations (3.77-4.28 mg/l) were measured from the acidic water layers at 22 m in the N pit.

Concentrations of soluble Fe were on average greater in waters of the Z than N pit (Table 6). In both pits surface and mid waters had markedly less soluble Fe than deeper waters, and the concentrations were slightly greater in March than in September. In the N pit, the Fe concentration ranged from 0.16-0.75 mg/l in surface and mid waters, and from 36.6-38.2 mg/l in deeper waters (≥35 m). In Z pit, it ranged from 0.05-1.15 mg/l in surface and mid waters, and from 5.12-411 mg/l in deeper waters (≥35 m). In both pits, the soluble iron was in reduced form, Fe²⁺. Unexpectedly, the Fe²⁺ concentrations were somewhat greater than the soluble Fe concentrations measured with ICP-OES. Obviously, some iron in the filtered water sample preserved with supra pure nitrogen acid has oxidized before the ICP-OES determination. The water samples for the Fe²⁺analysis were preserved with concentrated hydrogen chloride acid that obviously prevents Fe oxidation better than the nitrogen acid preservative.

Similarly to Fe, concentrations of soluble Mn were on average greater in the Z (7.79 mg/l) than N pit (1.12 mg/l). The concentrations were slightly greater in March than in September in both pits, except for the primary sample taken at the depth of 45 m from the Z pit in March (Table 6).

Metals associated with the Hammaslahti Cu-Zn ore had the greatest soluble concentrations in waters of the N pit, whereas their concentrations were pretty small, especially Cu, in waters of the Z pit (Table 6). Concentrations of Zn varied from 20-1,810 µg/l in the N pit and from 22-92 µg/l in the Z pit. Concentrations of Cu were much lower, varying between <0.1 µg/l and 60 µg/l in the N pit, and 0.2 µg/l in the Z pit. Similar to Fe and Mn, Zn concentrations were slightly greater in March than in September, whereas the relationship of Cu concentration with seasons was ambiguous.

Similar to Zn and Cu, concentrations of Ni and Co were on average greater in the N (Ni 54 µg/l, Co 10 µg/l) than Z pit (Ni 17 µg/l, Co 7 µg/l, Table 6). Waters of the N pit had somewhat lower concentrations of As in March than in September, whereas surface waters of the Z pit had no measurable As, and in deeper waters As ranged from 0.07-0.12 µg/l. Waters of the N pit had some soluble Pb(0.05-4.4 µg/l), whereas waters of the Z pit had Pb concentrations below the detection limit (≤0.05 µg/l). Concentrations of U in the majority of the pit water samples were less than 1 µg/l. Only deeper waters of the Z pit had some U (2.3-4.2 µg/l).

Table 6. Soluble concentrations of the essential major and trace elements of water samples taken from different depths in the N and Z pit, the closed Hammaslahti mine area, eastern Finland. Keys: < below the lowest detection limit, – no measurement due to interferences in the Fe²⁺ determination with spectrophotometer.

Water quality of the overflow of the Z pit and surface waters of the wetland pool complex

The pH of the overflow water was neutral (7.0), and close to neutral (at pH 7.2) at the depth of 2 m in the Z pit (Table 7a, see also Table 4). In the first wetland pool (Wetl1), it dropped slightly to 6.8 but rose to 7.2 in the second wetland pool (Wetl2). The pH of the water in the receiving ditch gradually decreased from 6.8 in ditch1 to 5.3 in ditch2. The redox potential was 170 mV in the overflow water, ranged from to 150-180 mV in the wetland pools, and markedly decreased (<60 mV) downstream in the receiving ditch, where water was more reducing than waters in the wetland pools. At the furthermost ditch sampling site redox potential was 130 mV. The decrease in the redox potential was linked to consumption of dissolved oxygen. The oxygen concentration was greater in surface waters (7.5-7.9 mg/l) of the wetland complex and the overflow water (6.6 mg/l) than in waters of the receiving ditch (3.7-5.8 mg/l).

Alkalinity of the discharge waters was 1.1 mmol/l. Alkalinity fluctuated from 0.9 to 1.0 mmol/l in the wetland pools and rose to 1.2 mmol/l at furthermost site of the receiving ditch (Table 7a).

Waters only had measureable solid particles in samples taken from the second wetland pool (3.3 mg/l), while abundant solid particles were measured in the furthermost site of the receiving ditch (46 mg/l) (Table 7a). The occurrence of solids presumably is due to a shallow depth of the water bed from which fine solids were easily incorporated during sampling with syringe.

According to the measurements, organic carbon was mainly insoluble, since the difference between the concentrations of unfiltered (total) and filtered (dissolved) samples, which indicates carbon bound by solid particles, was pretty low, ranging from 0.1-0.8 mg/l. Concentrations of the dissolved organic carbon gradually decreased from 3.0 mg/l in the overflow water to 1.6 mg/l in the receiving ditch water. Dissolved organic carbon increased to 5.0 mg/l in the ditch at the furthermost site (Table 7a). It is notable that dissolved organic carbon content decreases from the overflow point, downstream through the wetland pools. The organic carbon content rose at the furthermost sampling site, which is located in the bog, and where the water was assumedly mixed with peaty water.

The major anion was sulphate, of which concentration varied between 760 mg/l and 840 mg/l (Table 7a). It should be noticed that the amount of sulphate in waters slightly increased from overflow (760 mg/l) downstream into the wetland pools (ranging from 780 mg/l to 810 mg/l), and then slightly decreased (800 mg/l) at the first site of the receiving ditch, finally increasing to 840 mg/l. The comparison of the sulphur measurements from the IC (calculated as sulphate sulphur, S-SO₄) with those of the ICP-OES reveals that the amount of sulphate sulphur was somewhat greater in the water of the second wetland pool and in the water at the furthermost site of the receiving ditch than in the waters of the other sites. This may indicate the dissolution of hydrogen sulphide gas, or dissolved sulphide in waters of the second wetland pool and receiving ditch. This interpretation is based on the fact that the water sample for the IC measurements was not preserved. Therefore potential sulphide forms of sulphur can be oxidized to sulphate, whereas this is unlikely in the preserved filtered samples of the ICP-OES measurements.

In contrast to sulphur, concentrations of chloride stayed pretty unchanged from the overflow water (2.2 mg/l) to downstream in the wetland pools (2.2-2.3 mg/l) and receiving ditch (2.3-24. mg/l, Table 7a). A similar trend was also characteristic for F and NO₃ concentrations. Concentrations of F varied from 0.7-0.8 mg/l, and those of nitrate from 1.3-1.6 mg/l.

Concentrations of earth alkaline and alkaline metals varied a little, though the variation did not provide insight into their role in water remediation and buffering reactions (Table 7b). The soluble concentrations were 260-280 mg/l Ca, 39-44 mg/l Mg, 7.7-8.0 mg/l Na and 9.5-20 mg/l K. Concentrations of Al varied even less, staying around 0.2 mg/l. In contrast, concentrations of Fe, Mn and basic trace metals (Zn, Cu, Ni, Co) fluctuated between sampling sites. The iron concentration varied from 0.08-0.1 mg/l in waters of the overflow and first wetland pool. It rose to 1.6 mg/l in the water of the second wetland pool, dropped to <0.05 mg/l in the water of the receiving ditch, and rose again to 26.4 mg/l in the furthermost site of the receiving ditch. The majority of soluble Fe is in the form of reduced Fe²⁺. It is worthwhile to notice that the water of the second wetland pool and the furthermost receiving ditch site had somewhat greater concentrations of Fe²⁺ than soluble Fe. Assumedly, this indicates greater oxidation of Fe²⁺ in the samples preserved with supra pure nitric acid than in those preserved with hydrogen chloride acid.

Concentrations of Mn were lowest (0.1-0.2 mg/l) in waters of the outflow and the first wetland pool, but rose and remained constant (0.6 mg/l) in waters of the second wetland pool and receiving ditch (Table 7b). Similar to Fe, Mn concentrations markedly rose in the water at the furthermost site of the receiving ditch. Concentrations of soluble Zn, Cu, Ni and Co fluctuated from site to site, which indicates their instability in waters of the wetland pools and receiving ditch. In contrast, concentrations of Cd, As and Pb were below detection limits at most study sites. The exception was As, of which concentration was above detection limit (0.14 µg/l) at the furthermost site of the receiving ditch. Concentrations of soluble U were small, varying between 0.2 µg/l and 0.4 µg/l.

Table 7. Physical and chemical content of the overflow water from the Z pit, waters in the wetland pool complex (Wetl1, Wetl2) and water in the receiving bog ditch (Ditch1, Ditch2), the closed Hammaslahti mine area, eastern Finland.

Microbial composition in the pit waters

The bacterial community size in the water column of the two pits in March 2014 varied between 1.0 and 4.6 × 10⁷ ml^-1 of oxygenated water, but was approximately 10-fold lower in the oxygen depleted water (Fig. 8). However, in the Z pit at 35 m depth the concentration of bacteria was elevated. Elevated microbial concentration was possibly due to high TOC, DOC and Fe concentrations (Tables 5 and 6).

Figure 8. The concentration of bacteria ml^-1 in the water column on the N pit (left) and Z pit (right). The dotted line indicates the transition zone between oxygenated and oxygen depleted water.

The bacterial communities in the N and Z pits were very diverse and distinct from each other. Especially high diversity was seen at 40 m depth in the water of the N pit and in the sediment of the Z pit (Figs. 9 and 10). The bacterial community in the N pit was dominated by putative sulphur oxidizing and iron reducing Burkholderiales in the surface water (Fig. 9, Farkas et al. 2013). The Burkholderiales group can adsorb trace metals, e.g. Ni, as well as leach metals from solids. Another prominent betaproteobacterial group was the Methylophiliales, which use methanol and methylated small organic substrate for energy and carbon source (Doronina et al. 2014). This substrate may originate from the degradation of organic matter. Members of the uncultured and uncharacterized group SBla14 betaproteobacteria were also common at 22 m depth in the N pit water, but their ecological role is not clear. At 40 m depth 8-22% of the bacterial community consisted of epsilonproteobacterial Campylobacterales bacteria. These bacteria belong to a specific genus, the Sulfuricurvum, which oxidize sulphides and other reduced sulphur compounds, and fix carbon from CO₂ (Handley et al. 2014). These bacteria may also be able to oxidize Fe(II) (Chan et al. 2013). The water at 40 m depth contained between 6 and 8% SRB belonging mainly to the Desulfobacterales clade. The bacterial community in the sediment of the N pit belonged almost exclusively (86-93%) to the fermenting Clostridiales bacteria. This group is responsible for the degradation of organic materials, producing acetic and lactic acids, ethanol, and also CO₂ and H₂ gases which feed microbes. The group has no active role in the sulphate reduction.

Figure 9. The bacterial community of the N pit water and sediment resolved by high throughput amplicon sequencing. The bars indicate relative abundance of bacterial clades. Each sampled depth is presented with two replicate samples.

The dominating bacterial lineage in the Z pit water column belonged to the Gallionellales order. These bacteria are common iron oxidizers that precipitate iron as iron oxyhydroxide (Hallbeck et al. 1993). They can also oxidize Mn and As. The relative abundance of Gallionellales bacteria was highest in the water column at the oxycline, i.e. the transition zone between oxygenated and oxygen depleted water (Fig. 10). SRB belonging to deltaproteobacterial clades (Desulfuromonadales and Desulfobacterales) were present only at depths of 35 m and below, and were only below 1% of the total bacterial community. However, their abundance increased slightly in the sediment, where deltaproteobacterial SRB formed 1-2% of the bacterial community. Burkholderiales bacteria were a major group (14-16%) in the oxygenated water of the Z pit. Burkholderiales include species that oxidize sulphur compounds and reduce iron (Farkas et al. 2013). In addition, a great part of the bacterial community consisted of heterotrophic Bacteroidia bacteria (Saprospirales and Bacteroidales), which degrade organic matter and thus provide simple organic molecules for the benefit of the whole bacterial community (Fernández-Gómez et al. 2013). The Bacteroidales group was especially prominent in the sediment.

Figure 10. The bacterial community of the Z pit water and sediment resolved by high throughput amplicon sequencing. The bars indicate relative abundance of bacterial clades. Each sampled depth is presented with two replicate samples.

Mineralogical and chemical composition of the bottom sediments in the N and Z pits

According to XRD analysis, N-pit sediments contained both in March and September SiO₂ (Quartz) and FeS₂ (pyrite). In March, Al-silicates, chlorite, plagioclase and mica [(Mg, Al)₆(Si,Al)₄O₁₀(OH)₈, (Na,Ca)Al(Si,Al)₃O₈ and KMg₃(Si₃Al)O₁₀(OH)₂] were also found. In September, in addition to previously mentioned silicates [(Mg, Al)₆(Si,Al)₄O₁₀(OH)₈, (Na,Ca)Al(Si,Al)₃O₈, KAl₂(Si,Al)₄O₁₀(OH)₂)], there was K feldspar (KAlSi₃O₈).

According to XRD analysis, Z-pit sediments contained goethite (FeO(OH)), and some crystalline gypsum (Ca(SO₄)(H₂O)₂) in both March and September. Additionally, in September SiO₂, as amorphous silicon oxides (not crystalline quartz), was also identified.

The main elements of N-pit sediment were Si, Fe, Al and S, while the Z-pit sediments consisted mainly of Fe and in minor amounts S and Si bearing minerals (Table 8). In the N pit, the total S concentration of the sediment ranged from 6.1 w-% in March, to 7.5 w-% in September. The sediment of the Z pit had 1.0 w-% S in March and 1.4 w-% in September. Organic compounds were found only in the N-pit sediments, comprising 17% in March and 15% in September. There were some seasonal variations in the element contents which may be caused by poor sampling repeatability due to, for example, mineralogical and chemical heterogeneity of bottom sediments.

Table 8. Selected elements analyzed from pit sediments (XRF).

Organic compounds were found only from the N-pit sediments, 17% in March and 15% in September.

Nevertheless, the content of the acetate extractable fraction revealed the secondary dissolution and precipitation phenomena better than the total concentrations. The acetate extractable fraction represents easily leachable and exchangeable element concentrations (Räisänen at al. 1997, cf. Kumpulainen et al. 2007). Furthermore, the acetate extractable concentrations of Al and Fe may reflect their non-crystalline hydroxides, those of S to mono-sulphide, and obviously those of S and Ca to non-crystalline gypsum (CaSO₄).

The bottom sediment of the N pit was characterized by abundance of acetate extractable Fe, S, Ca and Al, while the sediment of the Z pit had the greatest concentrations of S and Ca, and less Fe and Al. It can be suggested that in the N pit, some Fe, S, Ca and Al were bound by non-crystalline precipitates (and perhaps also chemically adsorbed by organic compounds). Sources for the acetate extractable Fe are most likely non-crystalline Fe hydroxide and mono-Fe sulphide (cf. Kumpulainen et al. 2007). In the Z pit, S and Ca were the major elements in the sediment and can be easily mobilized under changing pH and redox conditions. The main source for acetate extractable S and Ca is non-crystalline gypsum.

Subtractions of the acetate extractable S and Fe concentrations from their total concentrations indicate the abundance of crystalline Fe sulphides in the N pit sediment. In Z pit, sulphur bound by the sediment was mostly acetate extractable. The quantity of acetate extractable Fe was very low compared to the total Fe concentration. These findings indicate a minor presence of Fe sulphide, but an abundance of crystalline Fe oxyhydroxides (cf. total and acetate extractable Fe concentrations).

Overall, the distribution of the acetate extractable elements in the sediments was ambiguous and showed some link to their soluble concentrations in overlying water (cf. Tables 6 and 9). The acetate extractable concentration of Fe was greater in September than in March in both pits. This indicates more release of Fe from sediments into the water column in March than in September, which approximately parallels the seasonal variation of soluble Fe in the water column, especially in the deep waters of both pits. In contrast, acetate extractable S seemed to be more adsorbed on solids and fixed with the secondary precipitates in March than in September in the N pit, whereas in the Z pit it behaved vice versa. However, these findings had no unambiguously link to seasonal S variations in the water column. It is noteworthy that Fe and S content in the duplicate water samples differed greatly, especially in March, which may result from the instability of the non-crystalline minerals.

Based on the acetate extractable concentrations, some of the trace metals (Zn, Ni, Pb) were adsorbed on sediment particles. However, their low concentrations obscure potential links to their soluble concentrations in the water column.

Table 9. Selected elements analysed with acid ammonium acetate (pH 4.5) extraction and ICP-OES.

Discussion

Water quality and acid generation in the N and Z pits

In both pits the physical and chemical quality of the surface water above 5 m was better than that of the mid and deep waters. An oxygen deficit occurred below the depth of 20 m in the N pit and below the depth of 30 m in the Z pit. The interface between oxic and anoxic water was about a meter higher in autumn than in spring. The present study showed no evidence of spring overturn of pit waters, but partial mixing of surface waters with mid waters occurred in the N pit. The acidification of oxygenated surface waters and especially water at the interface depths of 20 to 22 m, above the anoxic water, indicates the mixing of oxygen rich melt water downwards with relatively reduced waters in spring. Furthermore, the decrease in the pH results in dissolution and hydrolysis of Al at pH<5 that accelerates acidification. The enhanced acidity was followed by oxygen consumption in the water column that can be linked to the oxidation of soluble Fe²⁺ to Fe³⁺ (hydroxides/oxyhydroxides) coupled with the release of protons (Nordstrom & Alpers 1999, Bachman et al. 2001).

However, the predominance of ferrous Fe in the total soluble Fe content of the surface and acidified waters somewhat contradicted the above interpretation of the water acidification mechanism (cf. Ramstedt et al. 2003). Furthermore, the predominance of bacterial groups that degrade organic matter and potentially reduce Fe oxyhydroxides (consuming, not producing acid) suggest an alternate acidity source, such as the desulphurization of decaying organic material (e.g. chips) via photosynthetic oxidation in oxygenated waters (Sánches-España et al. 2008, NCSU 2015).This mechanism is supported by the discrepancy in total soluble sulphur and sulphate-sulphur results, which indicate the occurrence of reduced soluble sulphides potentially in both oxygenated and reduced water layers.

The profile of the bacterial community in the N pit water supports the reduction of Fe³⁺ in the water at least in the uppermost, oxygenated water layer where the Burkholderiales iron-reducing and sulphur-oxidizing bacteria dominated. This feature may also be present at 22 m depth, but the role of the betaproteobacterial SBla14 cluster is not yet clear. At greater depths Fe²⁺ is probably oxidized by epsilonproteobacteria. In addition, these epsilonproteobacteria detected in the N pit at 40 m depth are sulphur oxidizers, which could affect the acidification of water as well as prohibit the formation of Fe sulphides. SRB clades were generally detected only at relatively low abundances in the most diverse bacterial population at 40 m depth, and these bacteria coincided with the sulphur oxidizing epsilonproteobacteria. Thus, the results indicate that an active sulphur cycle is present at this depth. Chemical investigations indicated that sulphate reduction occurs in the anoxic parts of the pit. While the bacterial community reveals only low abundances of SRB in these parts of the pit, it is, however, possible that a small but efficient population of SRB is responsible for the formation of Fe sulphides.

The relatively high abundance of methylotrophic bacteria in the water column, especially at 22 m indicates high organic matter degradation with the production of methylated compounds and methanol. Methanol is produced by the degradation of, for example pectin, which is abundant in plant cell tissue. Such plant tissue material was added to the N pit in 2002 in the form of coniferous tree bark. Other sources of methanol and methylated compounds may include the pig manure additions (1998, 2000, 2004), hydrolysis of added absolute ethanol (2005) (Vestola and Mroueh, 2008), or recalcitrant carbon compounds trapped in the black shale of the rock. The high abundance of Clostridiales bacteria fermenting organic matter in the bottom sediments probably reflects the addition of high amounts of organic matter in the N pit. The fermentation processes release CO₂ and organic acids for use by other microbial groups. In addition, the fermentation process supports putative methanogenesis in the anoxic parts of the pit.

In contrast, the results from the Z pit waters showed no clear evidence of spring water overturn. The pH of surface and mid waters stayed pretty constant, near six (6.3-6.4), and the pH of deep waters varied a little from spring to autumn (from 6.2 to 6.5). The exception was the neutral pH of the topmost surface water in autumn. Obviously, this was caused by autumn rainfall. Similar increase in the autumn pH (but not as high as in the Z pit) characterized the topmost surface water in the N pit. On the basis of alkalinity, waters throughout the Z pit had excellent buffering capacity. Presumably, pit waters have been acidified, or are acidifying actively by Fe sulphide oxidation on pit walls. However, potentially because there was only a single sampling site in each pit, the results did not reveal active sulphide oxidation. Additionally, results do not show that acidic water from the N pit is flowing into the Z pit. The interpretation of active acid generation is based on the marked increase in alkalinity, especially in the deep waters of the Z pit.

In the Z pit the source for excellent buffering capacity (i.e. alkalinity) is obviously carbonate weathering followed by sulphide oxidation (of pit walls). According to Hämäläinen (1987), the Z ore deposit had calcite bearing veins of which weathering can explain additional release of soluble Ca in deep waters. However, crystalline gypsum (CaSO₄×2H₂O) was identified from the bottom sediments. Furthermore, the abundance of the acetate extractable S and Ca can indicate the precipitation of CaSO₄. The precipitation of gypsum is one of the main neutralizers of acidity. The bottom sediments were rich in Fe oxyhydroxides but no Fe sulphides was identified. This suggests that sulphate is first fixed with Fe oxyhydroxides, forming schwertmannite and/or ferrihydrite, which then transforms into crystalline goethite in deep waters and/or bottom sediments (Bigham et al. 1992, Kumpulainen et al. 2007, Sánchez-España et al. 2011). The transformation includes dissolution of sulphate and it can contribute to pH-buffering (Nordstrom & Alpers 1999). According to Bigham et al. (1992), the precipitation of Fe oxyhydroxides controls the formation of Fe sulphides. The above findings can explain the excess sulphur in relation to Ca and Fe in deeper waters. Furthermore, the abundance of soluble Mn may have restricted Fe sulphide formation in slightly acidic and reducing waters (Bigham et al. 1992, Kerrick & Horner 1998).

The high abundance of the iron oxidizing Gallionellales bacteria in the Z pit water strongly supports a biological iron oxidation process where iron is precipitated as Fe oxyhydroxides. In the oxygenated water from 1 to 30 m depth resultant Fe³⁺ may be reduced to Fe²⁺ by Burkholderiales bacteria, thus maintaining an iron reduction/oxidation cycle. The high diversity of bacterial groups present in the Z pit sediment indicates that the pit has not been influenced by anthropogenic organic carbon additions, as there is not a predominance of fermenting bacteria.

Water remediation capacity in the pits

The water quality of the samples taken in 2014 was compared to the data collected by Outokumpu Oyj in 2000 and 2001 after the addition of big manure and carbon source (Table 10). The monitoring data of Outokumpu Oyj included only a few parameters, i.e. pH, EC, Fe, Mn, Zn and SO₄, and therefore assessment of the rate and effectiveness of remediation can only be preliminary.

After around 15 years, the pH of waters in the N pit has decreased on average 20%, i.e. the acidity has increased about one pH-unit, whereas the pH of waters in the Z pit have increased on average 8%, i.e. 0.5 pH-units (Table 10). During the same period, the concentrations of the metals decreased variably from 50% to 80%. The retention of Fe and Mn was overall better in the N (80%) than Z pit (50-60%). Conversely, Zn concentration in solution increased on average over 90% in the N pit from 2000-2001 to 2014, while in the Z pit, it decreased 70%. The retention of SO₄ was much better in the N (70%) than Z pit (30%). The reason for this is obviously the notable formation of Fe sulphides in the deep waters of the N pit, whereas the formation of Fe oxyhydroxides and the abundance of soluble Mn have restricted sulphide formation in the Z pit.

Overall, it can be concluded that the sulphate reducing bacteria addition together with organic carbon feed (big mature, wood chips) has promoted Fe sulphide formation in the N pit. This results in a good remediation rate for Fe, SO₄ and Mn, but not for Zn, nor the other basic metals and Al. The increase in Zn concentration after the bacteria treatment is caused by the accelerated production of acidity, especially during spring water overturn. It can be concluded that the sulphate and iron reducing reactions do not increase the buffering capacity enough to neutralize the acidity during spring overturn. Furthermore, the rocks in the N pit, in contrast to the Z pit, do not bear carbonate minerals and such silicates that can increase buffering capacity via weathering.

The chemical difference between the N and Z pit waters suggests that there is no tunnel connection between the pits. In the Z pit, Fe and SO₄ were mostly precipitated as Fe oxyhydroxides, and the mineralogy of the bottom sediments did not reveal the occurrence of Fe sulphides, despite the fact that the S results revealed some soluble sulphide in waters. However, it is expected that crystalline goethite and other ferrihydroxides adsorbed trace metals, and therefore the retention rate for Zn and other metals was pretty good. However, the formation of Fe oxyhydroxides, as well as the buffering reactions resulting from carbonate weathering, increases the solubility of base cations (Ca²⁺, Mg²⁺, K⁺, Na⁺) and sulphate. Therefore the discharge from the Z pit is characterized by elements that increase the water salinity but not the content of trace metals downstream.

Table 10. Mean pH, electrical conductivity (EC), concentrations of Fe, Zn, Mn and SO₄ in surface, mid and deep waters of the N and Z pit in 2000-2001 and 2014, the closed Hammaslahti mine area, eastern Finland. The data of 2000 and 2001 is based on unpublished monitoring data of the Outokumpu Oyj.

Water remediation mechanisms in the N and Z pits

Boundary conditions for biological sulphate reduction in the bottom water layer of the mine pit are characterized by anoxic water conditions, organic substrate, and SO₄^2- (Lu 2004, Vestola & Mroueh 2008). Other recommended parameters are pH >5.5, Eh (redox) potential <-100 mV (García et al. 2001), and temperature >6 °C (Vestola and Mroueh 2008). When observing the N-pit (which has been the project site for biologic sulphate reduction) the bottom water was clearly anoxic in March, and almost completely anoxic in September, indicating suitable conditions for SRB. However, the observed redox-potentials were rather high for biologic sulphate reduction. SRB have the ability to generate negative potentials (Sheoran et al. 2010), but this has not happened during the 15 years of treatment in the N-pit. However, SRB can also perform successfully under positive Eh conditions when anoxic, reducing “pockets” in organic substrate exist (Sheoran et al. 2010). It is very possible that these pockets are found in the deepest water, just above the bottom sediments, but unfortunately microbiological sampling found only minimal SRB in the sediments. In addition to slightly unfavourable Eh potential, the temperature in the bottom water is very low for SRB, which greatly decreases their activity. The pH is suitable for sulphate reducing bacteria, as well as the concentration of SO₄^2- and organic substrate in sediments.

The N pit sediment samples contained high concentrations of Fe, S and elevated Zn, all of which were also elevated in the overlying water. It seems clear that some biologic sulphate reduction is occurring in the bottom of the N pit, as FeS₂ was the only species containing Fe and S according to XRD analysis. In addition, SO₄ and Fe concentrations have clearly decreased during the treatment (Fig. 11). Therefore, it is estimated that SRB are reducing SO₄ from the bottom waters to hydrogen sulphide, which then precipitates Fe²⁺ to FeS (Sheoran et al. 2010):

3SO₄²⁻ + 2CH₃CHOHCOOH → 3H₂S + 6HCO₃⁻

H₂S + Fe²⁺ → FeS(s) + 2H⁺

Another possible reaction for Fe²⁺ to FeS is (Sheoran et al. 2010):

CH₂O + 4FeOOH + 8H⁺ → CO₂ + 4Fe²⁺ + 7H₂O

2FeOOH + 3H₂S → 2FeS + S⁰ + 4H₂O

The FeS product then reacts to insoluble pyrite (FeS₂) via various reactions (Vestola and Mroueh, 2008):

FeS + S⁰ → FeS₂

FeS + S_x^2- → FeS₂ + S_(x-1)^2-

FeS + H₂S → FeS₂ + H₂

Zn was detected in N pit sediments (XRF), but no Zn compounds were found by XRD. According to XRF (Table 8) and ammonium acetate extraction (Table 9), only 2-6% of Zn was easily leachable, indicating that the majority of Zn was precipitated as ZnS by biologically produced H₂S.

H₂S + Zn²⁺ → ZnS(s) + 2H⁺

The stability of the subsequent precipitate plays an important role when treating sulphate and metal laden waters. In biologic sulphate reduction, bacteria oxidize organic substrate and reduce SO₄^2-, producing H₂S or HS^–, which readily precipitate iron and various metals as insoluble sulphides (Cocos et al. 2002). However, in this study only FeS₂ was found in sediments. Therefore stability was studied by estimating the amount of easily soluble metal compounds with the acid ammonium acetate method and comparing obtained values with total concentrations found in sediments by XRF. This comparison (Fig. 11) showed that all studied metals were occurring largely as insoluble precipitates, and the share of soluble species was ≤ 22% for Fe, ≤ 14% for Zn, Mn and Ni, and ≤ 10% for S. In March, the level of all metals was higher (both easily soluble and total concentration) than in September. The situation was opposite for sulphur, as its concentration increased during the summer.

Figure 11. Total concentration of metals (red column) compared to easily soluble part (blue column) in March and September, N-pit.

It is estimated that a share of non-easily soluble sulphur is occurring as metal sulphides in the bottom sediments. During the year, pH and redox in the bottom water were 5.6 to 5.7 and -4 to-58 mV, respectively. From the Figure 12 it can be estimated that Cu, especially, is completely insoluble, and Zn only slightly soluble. The estimations of Figure 12 align very well with observations from the N pit bottom water (Table 6).

Figure 12. Soluble concentration of metal sulphides (Peters & Ku 1985, Hammack et al. 1994).

The stability of FeS₂ is not clear in the N pit. It has been calculated that in reductive marine waters, FeS₂ requires clearly negative Eh potential to be stable. In the Hammaslahti N pit, the redox potential ranged from -58 mV to 0 mV when measured with Ag/AgCl electrode, equivalent to values between +140 mV and 200 mV (SHE). As shown in Figure 13, FeS₂ tends to dissociate to Fe²⁺ and SO₄ under these conditions (Glasby & Schultz 1999). According to our highly simplified models (which include corrected pressure, temperature and concentrations of the main dissolved elements for the N pit), the Fe solution chemistry is not as clear as in Figure 13, but the modelled results still illustrate that FeS₂ may be unstable, especially during the summer when overturn can increase redox potential. However, biologic sulphate reduction seems to be more active than the dissolution of FeS₂, as concentrations of SO₄ and Fe have clearly decreased during the treatment (Fig. 11).

Figure 13. Eh-pH diagrams of the aquatic and solid phases of Fe, calculated from deep-seawater conditions in the Angola Basin. Temperature 2 ^oC, pressure 1 atm, Eh as SHE (Glasby and Schultz 1999).

Water remediation in the wetland pool complex

The remediation rate in the wetland pool complex that receives overflow waters from the Z pit is estimated by comparing water quality data from 2000 and 2001 collected by the GTK to the data of the present study. The comparable data from the years of 2000 and 2001 includes samples from the surface pit water before discharging (the current overflow), the ditch water receiving the overflow in 2000 and 2001 (the current ditch1), and samples taken close to the furthermost site of the ditch (ditch2, See Räisänen et al. 2003).

The most notable impact of the wetland pool complex on the water quality over 15 years is the decrease in water acidity (Table 11). However, the increase in pH was not as marked at the furthermost ditch site. The greater acidity downstream of the ditch is assumed to be caused by other acidity sources than the discharge water from the wetland complex. The ditch2 site (the peat) was earlier contaminated not only by acidic discharges directly from the northern corner of the Z pit, but also from seepage waters from the tailings facility.

In addition, the amount of base cations and sulphate (sulphur in Table 11) has decreased 40 to 50% (Fig.14). The most notable change is for Mn and other trace elements, of which concentrations have decreased over 90%. The concentration of Fe dropped almost 100% in the overflow waters, but the amount of soluble Fe increased about 20% at the furthermost ditch site. This is assumed to be caused by the earlier contamination of the peat close to the ditch.

The difference between the pH means of the surface waters of the Z pit in Table 8 and those of the overflow water in Table 11 in 2000-2001 is caused by the differences in sampling dates. The pit waters collected by Outokumpu Oyj were taken only two times in March and once in August, whereas the GTK collected samples both in 2000 and 2001 in May, June, August and October. In the GTK’s data, the pH of the surface water was usually acidic (≤4) in summer months, lowering the mean values in Table 11. In addition to the current pH data, the monitoring pH data from 2012 confirms that the pH of the wetland pool waters has stabilized between 6 and 7 (Viitasalo 2013). Moreover, the content of base cations and sulphate has become stable, staying unchanged throughout the wetland pool complex, downstream to the furthermost site of the receiving ditch. Aluminium had a similar distribution trend. Concentrations of Fe and several trace metals (Mn, Zn, Cu, Ni, Co) fluctuated in minor amounts from the overflow to the wetland pools and receiving ditch without fast retention in wetland or ditch sediments. Therefore, it is concluded that water remediation occurred largely inside the pit. The wetland pool complex and receiving ditch water quality changes are more dependent on discharges from the tailings facility and the stability of contaminants trapped in the wetland and ditch sediments, than on the overflow water quality.

Table 11. Physical and chemical quality of the overflow water of the Z pit and that of the receiving ditch water in 2000-2001 and 2014, the closed Hammaslahti mine area, eastern Finland (See sites in Fig. 4). Key: n refers to the amount of water samples.

Figure 14. Decrease (%) in the element concentrations of the overflow water of the Z pit and the receiving ditch water from years 2000-2001 to the year of 2014, the closed Hammaslahti mine area, eastern Finland. (See sites in Fig. 4).

Conclusions

The present study revealed marked differences in physical, chemical and biological contents of the waters in two open pits (N, Z) in the closed Hammaslahti mine. This indicates that there is no tunnel connection between the N and Z pits. The physical and chemical quality of the surface water above 5 m in both pits was better than that of the mid and deep waters. An oxygen deficit occurred below 20 m in the N pit and below 30 m in the Z pit in spring as well as in autumn. Deep waters and bottom sediments of both pits were characterized by year-round slightly reducing and pH≥5.5 conditions.

After around 15 years, the pH of waters in the N pit had decreased on average 20%, whereas the pH of waters in the Z pit had increased on average 8%. During the same period, the concentrations of the metals decreased variably from 50% to 80%. The retention of Fe and Mn was overall better in the N (80%) than Z pit (50-60%). Also the retention of SO₄ was much better in the N (70%) than Z pit (30%).

The chemical content of the overflow from the Z pit stayed pretty unchanged throughout the wetland pool complex, downstream to the furthermost site of the receiving ditch. The wetland complex rich in cattails and reeds maintains neutral and slightly oxidizing (water) to reducing (sediment) conditions that promote biological sulphate reducing and metal retention in organic substrate. Overall, the pit waters are mainly remediated in the pits, and the wetland complex had minor influence on the composition of overflow water. Seepages from the tailings impoundment notably deteriorate the discharge from the wetland complex downstream in bog drains.

Bacterial communities in the N and Z pits were very diverse and distinct from each other. In the oxygenated waters of both pits, the sulphur reducing and iron oxidizing Burkholderiales bacteria were common, whereas the deep water and bottom sediments contain a small but efficient population of sulphate reducing bacteria (SRB) promoting the formation of Fe sulphides. However, SRB groups coincided with sulphur oxidizing epsilonproteobacteria that resulted in an active sulphur cycle and moderately high solubility of sulphate in deep waters. On the other hand, the cold water temperature (<+6 °C) and slightly unfavourable redox potential (>-100 mV) are obvious reasons for a great decrease in SRB activity.

In the Z pit, Fe and SO₄ were mostly precipitated as Fe oxyhydroxides. The mineralogy of the bottom sediments showed no Fe sulphides despite the fact that the geochemical fractionation of S showed some soluble sulphide in deep waters that had minimal SRB. Crystalline goethite and other ferrioxyhydroxides adsorbed sulphate and trace metals. Therefore water remediation can be considered fairly effective. Furthermore, buffering reactions largely driven by carbonate weathering on pit walls resulted in gypsum precipitation and sufficient neutralization of the pit water overflowing into the wetland complex.

In contrast to waters of Z pit, the oxygenated water of the N pit showed acidification, most perceptibly, as a result of spring overturns. Despite Fe²⁺ oxidation caused by oxygen rich melt water and following by the release of Al and protons, it is assumed that acidity in the N pit is released by desulphurization of decaying organic material (e.g. wood chips) via photosynthetic oxidation.

The Hammaslahti N-pit was assessed to be a challenging site for biologic sulphate reduction based in-situ pit treatment due to low temperatures. However, SO₄ and Fe concentrations in the bottom water have decreased. It is hypothesized that the successful campaign of organic substrate additions has promoted the activity of sulphate reducing bacteria, which have converted SO₄ to H₂S (HS^–). The hydrogen sulphides have precipitated iron and zinc as FeS₂ and ZnS. Unfortunately, the pH of the system has not increased to neutral, possibly due to insufficient activity or abundance of sulphate reducing bacteria. Indeed, very low amounts of SRB were detected from the sediments and waters during sampling. This is most likely due to the low temperature of the N pit waters. The low activity of SRB is a drawback to the system, as strongly reducing conditions are not induced, making the dissolution of FeS₂ possible; the ZnS seems to be more stabile. Improvements to the N pit biological treatment can most likely be achieved by the addition of organic substrate in the future to promote the development of the sulphate reducing bacteria community. One possibility would also be to introduce more cold-tolerant bacteria to the pit.

Performance

Water treatment in two open pits (N and Z) in the closed Hammaslahti mine area is based on passive biogeochemical retention of metals and sulphur. The overflow from the northernmost (Z) pit is discharged into the partially constructed wetland pool complex where the passive remediation of water is expected to continue. To promote biologic sulphate reduction, pig manure (sludge) as a bacteria matrix, was added three times (1998, 2000 and 2004) and followed by additions of wood park chips (2002) and ethanol (2005) that decay slowly and release carbon as an energy source for bacteria. According to water and sediment analysis results, water remediation in the N pit is mainly driven by metal sulphide formation. In the Z pit the formation of Fe oxyhydroxides and their adsorption capacity facilitate metals retention. Despite finding SRB in the deep water of the Z pit, the bottom sediments rich in goethite and gypsum had no sulphide minerals.

Originally the N pit was the pilot site for biological sulphate reduction and it was presumed that sulphate reducing pockets within organic substrate would spread via a tunnel to the Z pit. Notable differences in geochemistry and the microbial content of waters and bottom sediment of the two pits suggest there is not a hydraulic connection via a tunnel between the N and Z pit.

Based on the pit water results, the remediation capacity can be rated good and the self perpetuating remediation mechanisms in both pits facilitate passive remediation. However, it is recommended to collect biological and geochemical data from bottom sediments from different points of both pits to increase understanding of cold water microbial activity and its impact on water remediation.

Results from the wetland pool complex showed that the overflow water from the Z pit remained pretty unchanged throughout the wetland complex and notably, deteriorated downstream after mixing with seepage waters from the tailings impoundment. Therefore, the remediation mechanisms in the Z pit are the main water quality control of the pit water. The neutral and near-reducing conditions of the wetland complex may act to reduce sulphate, retain chalcophile metals, and provide neutralizing capacity to the system in the event of overflow acidification.

The Hammaslahti N pit case is an example of biologic sulphate reduction, whereas the Z pit case is an example on Fe oxidation coincident with gypsum precipitation. This study was carried out with geochemical and biological analyses from minimal samples. Regardless, it provided insight to the physical and chemical quality of the pit waters and the main remediation mechanisms. Some drawbacks were identified. Despite the additions of pig manure bacteria matrix and organic substrate energy source material, strong reducing conditions were not induced. One reason for inadequate activity of the SRB could be the cold temperature of the pit water, especially at the bottom of the pit. This suggests the need for further study of the behaviour of cold-tolerant bacteria and their impact on water remediation. Results from the Z pit revealed that both oxidizing and reducing bacteria can maintain reactions that retain sulphate and metals via accumulation of stable secondary precipitates at the bottom of the pit.

Due to limited data collected during the case study, it was not possible to estimate the actual cost of the passive treatment technology carried out in the Hammaslahti closed mine area.

References

Bachman, T.M., Friese, K. & Zachmann, D.W. 2001. Redox and pH cpnditiond in the water column and in the sediments of an acidic mining lake. Geochemical Exploration 73, 75-86.

Bigham, J.M., Schwertmann, U. & Carlson, L. 1992. Mineralogy of precipitates formed by biochemical oxidation of Fe(II) in mine drainage. Teoksessa: H.C.W. Skinner & R.W. Fitzpatrick (eds.) Biomineralization – Processes of Iron and Manganese – Modern and Ancient Environments. Catena Verlag, Cremlingen-Destedt, Catena Supplement 21, 219-232.

Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E. K. & Knight, R. 2010. QIIME allows analysis of high-throughput community sequencing data. Nature methods, 7, 335-336.

Carcia, G.P., Moreno, D A., Ballaster, A., Blásquez, M.L. & González, F. 2001. Bioremediation of an industrial acid mine water by metal-tolerant sulphate-reducing bacteria. Minerals Engeneering 9, 997-1008.

Chan, C.S., McAllister, S., Krepski, S., Lin, C., Lazareva, O. & Kan, J. 2013. A novel Fe (II)-oxidizing Epsilonproteobacterium from a streambank aquifer. In: AGU Fall Meeting Abstracts, vol. 1, 0377.

Cocos, I.A., Zagury, G.J., Clément, B. & Samson, R. 2002. Multiple factor design for reactive mixture selection for use in reactive walls in mine drainage treatment. Water Research 32, 167-177.

DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K., Huber, T., Dalevi, D., Hu, P. & Andersen, G.L. 2006. Greengenes, a Chimera-Checked 16S rRNA Gene Database and Workbench Compatible with ARB. Applied and Environmental Microbiology 72, 5069-72.

Doronina, N., Kaparullina, E. & Trotsenko, Y. 2014. The Family Methylophilaceae. In: The Prokaryotes: Alphaproteobacteria and Betaproteobacteria pp. 869-880.

Farkas, A., Dragan-Bularda, M., Muntean, V., Ciataras, D. & Tigan, S. 2013. Microbial activity in drinking water-associated biofilms. Central European Journal of Biology 8, 201-214.

Fernández-Gómez, B., Richter , M., Schüler, M., Pinhassi, J., Acinas, S. G., González, J.M. & Pedrós-Alió, C. 2013. Ecology of marine Bacteroidetes: a comparative genomics approach. The ISME journal 7, 1026-1037.

García, C., Moreno, D.A., Ballester, A., Blásquez, M.L. & González, F. 2001. Bioremediation of an industrial acid mine water by metal-tolerant sulphate-reducing bacteria. Minerals Engineering 9, 997-1008.

Glasby, G.P. & Schultz, H.D. 1999. Eh, pH Diagrams for Mn, Fe, Co, Ni, Cu and As under Seawater Conditions: Application of Two New Types of Eh, pH diagrams to the Study of Specific Problems in Marine Geochemistry. Aquatic Geochemistry 5, 227-248.

Hallbeck, L., Ståhl, F. & Pedersen, K. 1993. Phytogeny and phenotypic characterization of the stalk-forming and iron-oxidizing bacterium Gallionella ferruginea. Journal of General Microbiology 139, 1531-1535.

Hammack, R.W., Edenborn, H.M. & Dvorak, D.H. 1994. Treatment of water from an open-pit copper mine using biogenic sulfide and limestone: a feasibility study. Water Research 28, s. 2321–2329.

Hämäläinen, I. 1987. Hammaslahden sinkkimalmi. Julkaisematon Pro gradu-tutkielma. Turun yliopisto, Geologian ja mineralogian laitos. 110 s.

Handley, K.M., Bartels, D., O’Loughlin, E.J., Williams, K.H., Trimble, W.L., Skinner, K. & Gilbert, J.A. 2014. The complete genome sequence for putative H2‐and S‐oxidizer Candidatus Sulfuricurvum sp., assembled de novo from an aquifer‐derived metagenome. Environmental microbiology 16, 3443-3462.

Karppanen, T. 1986. The Hammaslahti Mine. In: Kojonen, K. (ed.) Prospecting in areas of glaciated terrain 1986: guide to the field excursion for the symposium in eastern Finland, September 3-5 1986, 34-40.

Kennedy, A. 1985. Mineral processing developments at Hammaslahti, Finland. Mining Magazine-February 1985, 122-129.

Kerrick, K H. & Horner, M. 1998. Retention of manganese by a constructed wetland treating drainage from a coal ash disposal site. Proceedings of the annual meeting. American Society for Surface Mining and Reclamation 15, 272-279.

Kumpulainen, S., Carlson, L. & Räisänen, M. L. 2007. Seasonal variations of ochreous precipitates in mine effluents in Finland. Applied Geochemistry 22, 760-777.

Loukola-Ruskeeniemi, K., Gaal, G. & Karppanen, T. 1992. Geochemical and structural characteristics of a sediment-hosted copper deposit at Hammaslahti: comparison with Beshhi-type massive sulphide deposits. In: K. Loukola-Ruskeeniemi, Geochemistry of Proterozoic metamorphosed black shales in eastern Finland, with implication for exploration and environmental studies. Academic dissertation, University of Helsinki, Finland. Paper III.

Lu, M. 2004. Pit lakes from sulphide ore mining, geochemical and limnological characterization before treatment, after liming and sewage sludge treatments. Doctoral Thesis, Luleå University of Technology.

NCSU 2015. The sulphur circle. http://www4.ncsu.edu/~franzen/public_html/Poland/Poznan08a/Sulfur_Cycle.pdf Date 23042015

Nordstrom, D.K. & Alpers, C. N. 1999. Geochemistry of acid mine waters. In: G.S. Plumlee & M. J. Logsden (eds.) The Environmental Geochemistry of Mineral Deposits, Part A. Processes, Techniques, and Health Issues. Society of Economic Geologists, Revision Economy Geology 6 A, 133-156.

Pelkonen, K., Alopeus, E. & Penttilä, S. 1973. Outokumpu Oy:n Hammaslahden kaivos. Vuoriteollisuus 31 (2), 90-96.

Peters, R.E. & Ku, Y. 1985. Batch precipitation studies for heavy metal removal by sulfide precipitaion. AIChE Symposium Series 81, s. 9–27.

Puustinen, K. 2003. Suomen kaivosteollisuus ja mineraalisten raaka-aineiden tuotanto vuosina 1530–2001, historiallinen katsaus erityisesti tuotantolukujen valossa, Geologian tutkimuskeskus, arkistoraportti, M 10.1/2003/3. 578 s, http://weppi.gtk.fi/aineistot/kaivosteollisuus/.

Räisänen, M L. 2003. Rehabilitation options for tailings impoundments – case studies of wet cover and wetland treatment. In: C. Hebestreit, J. Kudełko and J. Kulczycka (eds.) Mine Waste management Best Available Techniques. CBPM Cuprum, Wroclaw and MEERI PAS, Kraków, 141-150.

Räisänen, M.L., Kashulina, G. & Bogatyrev, I. 1997. Mobility and retention of heavy metals, arsenic and sulphur in podzols at eight locations in northern Finland and Norway and the western half of the Russian Kola Peninsula. Journal of Geochemical Exploration 59, 175-195.

Räisänen, M.L., Niemelä, K. & Saarelainen, J. 2003. Rautasulfidipitoisen rikastushiekan läjitysalueen rakenne ja ympäristön pintavesien nykytila. Vuosien 2000 ja 2001 seurantatulokset, Hammaslahden vanha kuparikaivos. Geologian tutkimuskeskus, arkistoraportti S/44/0000/1/2003. 27 s.

Ramstedt, M., Calrsson, E. & Lövgren, L. 2003. Aqueous Geochemistry in the Udden pit lake, northern Sweden. Applied Geochemistry 18, 97-108.

Sánchez-España, J., Pamo, E. L., Pastor, E.S. & Marta Diez Ercilla, M D. 2008. The acidic mine pit lakes of the Iberian Pyrite Belt: An approach to their physical limnology and hydrogeochemistry. Applied Geochemistry 23, 1260–1287.

Sánchez-España, J., Yusta, I. & Diez-Ercilla, M. 2011. Schwertmannite and hydrobasaluminite: A re-evaluation of their solubility and control on the iron and aluminium concentration in acidic pit lakes. Applied Geochemistry 26, 1752-1774.

Sheoran, A.S., Sheoran, V. & Choudhary, R.P. 2010. Bioremediation of acid-rock drainage by sulphate-reducing prokaryotes: A review. Minerals Engineering 23, 1073-1100.

Tenhola, M. & Räisänen, M.L. 2006. Rikastushiekan läjitysalueen suotovesien vaiku­tus ympäröivän turpeen kemialliseen koostumukseen, vanha Hammaslahden Cu-Zn -kaivos, Pyhäselkä. Geologian tutkimuskeskus, arkistoraportti S44/0000/1/2006. 28 s.

Tsitko, I., Lusa, M., Lehto, J., Parviainen, L., Ikonen, A.T.K., Lahdenperä, A.-M. & Bomberg, M. 2014. The variation of microbial communities in a depth profile of an acidic, nutrient-poor boreal bog in southwestern Finland. Open Journal of Ecology 4, 832-859.

Vestola, E. & Mroueh, U.-M. 2008. In Situ Treatment of Acid Mine Drainage by Sulphate Reducing Bacteria – Guide to the pit lake treatment. VTT Research Notes 2422 (in Finnish).

Viitasalo, M. 2013. Hammaslahden kaivosalueen kuormitus- ja vesistötarkkailun vuosiyhteenveto 2012. Savo-Karjalan ympäristötutkimus Oy E4324. 31 s.