Journal of Environmental Management 92 (2011) 2419e2426

Contents lists available at ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Bioremediation of Acidic and Metalliferous Drainage (AMD) through organic carbon amendment by municipal sewage and green waste
Clint D. McCullough*, Mark A. Lund
Mine Water and Environment Research Centre (MiWER), Edith Cowan University, 270 Joondalup Drive, Perth, WA 6027, Australia

a r t i c l e i n f o
Article history: Received 9 December 2010 Received in revised form 11 April 2011 Accepted 25 April 2011 Available online 26 May 2011 Keywords: AMD Bioremediation Sulphate reduction Pit lakes Organic amendments

a b s t r a c t
Pit lakes (abandoned ooded mine pits) represent a potentially valuable water resource in hot arid regions. However, pit lake water is often characterised by low pH with high dissolved metal concentrations resulting from Acidic and Metalliferous Drainage (AMD). Addition of organic matter to pit lakes to enhance microbial sulphate reduction is a potential cost effective remediation strategy. However, cost and availability of suitable organic substrates are often limiting. Nevertheless, large quantities of sewage and green waste (organic garden waste) are often available at mine sites from nearby service towns. We treated AMD pit lake water (pH 2.4) from tropical, North Queensland, Australia, with primarytreated sewage sludge, green waste, and a mixture of sewage and green waste (1:1) in a controlled microcosm experiment (4.5 L). Treatments were assessed at two different rates of organic loading of 16:1 and 32:1 pit water:organic matter by mass. Combined green waste and sewage treatment was the optimal treatment with water pH increased to 5.5 in only 145 days with decreases of dissolved metal concentrations. Results indicated that green waste was a key component in the pH increase and concomitant heavy metal removal. Water quality remediation was primarily due to microbially-mediated sulphate reduction. The net result of this process was removal of sulphate and metal solutes to sediment mainly as monosuldes. During the treatment process NH3 and H2S gases were produced, albeit at below concentrations of concern. Total coliforms were abundant in all green waste-treatments, however, faecal coliforms were absent from all treatments. This study demonstrates addition of low-grade organic materials has promise for bioremediation of acidic waters and warrants further experimental investigation into feasibility at higher scales of application such as pit lakes. Crown Copyright 2011 Published by Elsevier Ltd. All rights reserved.

1. Introduction Large-scale open-cut mining activity has left a legacy of many thousands of mine pit lakes worldwide (Klapper and Geller, 2002). Pit lakes form from these voids when dewatering ceases in a mine pit that extends below the water table (Castro and Moore, 1997). As groundwater, and surface runoff slowly ll the pit, this water may react with seams and pit walls resulting in dissolution and oxidation of minerals resulting in acidication and metal contamination of contacting waters as Acidic and Metalliferous Drainage (AMD) (Castro and Moore, 1997). AMD is arguably the greatest environmental problem facing water management in the international mining industry (Gray, 1997). AMD degraded environments such as pit lakes typically have reduced environmental and social values, and the resultant water is less valuable as a resource to adjacent mine operations and regional communities (Doup and Lymbery, 2005). Such contaminated water bodies may also directly or indirectly threaten the health and well-being of
* Corresponding author. Tel.: 61 8 6304 5143; fax: 61 8 6304 5509. E-mail address: c.mccullough@ecu.edu.au (C.D. McCullough).

local communities and natural ecosystems (McCullough and Lund, 2006). Nevertheless, the vast quantities of water in many pit lakes (often billions of litres) may also represent potentially valuable resources to mining companies, the environment and community, if appropriate water quality can be achieved (McCullough et al., 2009). In situ microbial sulphate reduction is a developing ecotechnology with potential as an efcient and effective approach for remediation of the vast volumes typically found in acidic pit lakes (Wielinga, 2009). In situ neutralisation via sulphate reduction with amended organic matter is expected to play a key-role in the remediation of acidic mining pit lakes (Kleeberg, 1998). Sulphate reducing bacteria (SRB) can produce alkalinity by converting sulphate to sulde when supplied with labile carbon sources in low redox environments (Frmmichen et al., 2004; Wendt-Potthoff et al., 2002). Dissolved metals can then bind with sulde to form insoluble precipitates thereby removing metals and increasing pH (Ksel and Dorsch, 2000; McCullough, 2007; Praharaj and Fortin, 2004). The composition of organic matter has been found to be an important determinate of SRB ecotechnology effectiveness, as it

0301-4797/$ e see front matter Crown Copyright 2011 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvman.2011.04.011

2420

C.D. McCullough, M.A. Lund / Journal of Environmental Management 92 (2011) 2419e2426

must both create a suitably low redox environment and serve as readily-available carbon source (Gibert et al., 2002). For example, studies with various plant matter as organic amendments have found that availability of carbon is dependent upon the rate of decomposition, which can be extremely limited in acidic and anoxic conditions (Neculita et al., 2007). As a result, most published studies into the feasibility of sulphate reduction based bioremediation have focused on highly labile carbon substrates, yet with limited availability in many mining areas, such as ethanol (Kolmert and Johnson, 2001), sugar and alkaline sugar production byproducts (Geller et al., 2009), wine and cheese industry waste products (Martins et al., 2009) and animal manure (Drury, 1999). Nevertheless, viability of sulphate reduction based bioremediation requires an organic amendment that is effective, yet also economical and locally available in bulk (Castro and Moore, 1997; Kumar et al., in press-a). In many remote mining locations, the only bulk organic carbon sources likely to meet such criteria are sewage (from the site or support town) and green waste (a broad range of plant material collected through clearing of areas on the mine or in the towns from domestic and municipal lawns and gardens) (Kumar et al., in press-b). This paper reports on a laboratory scale microcosm study of the suitability of green waste and sewage as an organic matter source for bioremediation of AMD contaminated pit lake water by microbially-mediated sulphate reduction. The experiment builds on the preliminary ndings using organic amendments in highly acidic AMD (McCullough et al., 2008a, 2006) by varying organic amendment dose under replicated and controlled experimental conditions and with greater elucidation of microbiological mechanisms. 2. Methods 2.1. Study area Collinsville is a small (population 2000) coal mining town, located 70 km from the coast in North Queensland, Australia (20 330 5300 S, 147 490 5200 E) with a semi-arid climate with a landscape of woodland-savannah. Precipitation is dominated by moderately low and episodic summer rainfall (annual mean of 708 mm) and very high annual evaporation rate (annual mean of 1860 mm). The Wet season (DecembereApril) temperature mean is 32.5  C and the Dry season (MayeNovember) temperature mean 28.5  C (Commonwealth of Australia Bureau of Meteorology, 09/02/2009). Typical of many mining operations in low rainfall regions (e.g., Southern Africa, Hobbs et al., 2008) water security is a priority concern for mining operations located in inland tropical Queensland (Australia). For example, there are 20 pit lakes on the Collinsville Coal Project (CCP) operational area in the North Bowen coal eld, containing in total at least 2 109e3 109 L, all of which are of extremely low pH (ca. pH 2) and contain high concentrations of solutes (electrical conductivity 9e19 mS cm1) and high OxidationeReduction-Potential (ORP, 640e560 mV, platinum reference electrode) (McCullough et al., 2006). Metal concentrations are also high throughout these pit lakes, although total nitrogen concentrations of 1 mg L1 are typically very low and Dissolved Organic Carbon (DOC) is moderate (13e52 mg L1) (McCullough et al., 2008b). Remediation of acidic mine lake water, even to only marginal quality, would reduce pressure on scarce regional water resources for low-quality industrial uses such as dust suppression or ore processing. 2.2. Experimental design Microcosms were constructed from clean 100 mm diameter and 600 mm long (4.5 L) acrylic tubes, sealed at the base with a rubber

Fig. 1. Microcosm design showing placing of sediment, organic matter and mesh tube for a GS (combined green waste and sewage) treatment and relative heights of amendments in microcosms.

bung and at the top with a loose-tting PVC cap for limited gas exchange between water and atmosphere (Fig. 1). Pit lake AMD water and sediment were collected from Garrick East pit lake (GAE), a typical North Bowen Coal Basin pit lake (McCullough et al., 2008b), on 30/06/2005. Microcosms were lled with lake littoral sediment to 140 mm depth and then GAE water to within 20 mm of the core top. A 10 mm diameter breglass mesh (ywire) tube was placed in each microcosm to allow measurement of pH from nearsediment waters through the amended organic matter layer. The lower 300 mm of all microcosms were placed in an opaque black plastic tub which was lled with tap water to evenly distribute ambient temperatures between cores and to reduce the possibility of leakage. Tubs were then wrapped to the top of the microcosms in black PVC sheeting to occlude light, as ambient light levels are extremely low in the hypolimnion of these lakes (authors unpublished data). 2.3. Organic amendment Primary-treated sewage sludge was collected from the drying beds of Collinsville and Bowen Municipal Wastewater Treatment Plants. Due to import restrictions between states, representative green waste was prepared from typical Perth (Western Australia) urban garden waste material by chopping into 50 mm lengths and drying at 25  C for one week to simulate Collinsville municipal green waste conditions. Homogenised samples of sewage and green waste were pH paste-tested (AR grade water:solid mass ratio of 10:1). Using organic substrates similar to McCullough et al. (2006) for a total of 21 microcosms, three random microcosms were untreated as controls, while treatments of three random replicate microcosms were treated as follows: a) 100 g green waste (G100), b) 200 g green waste (G200), c) 100 g of green waste and sewage combined (1:1) (GS100), d) 200 g of green waste and sewage combined (1:1) (GS200), e) 100 g sewage (S100), and f) 200 g sewage (S200). The 100 g treatments had a 32:1 water:organic mass ratio with

C.D. McCullough, M.A. Lund / Journal of Environmental Management 92 (2011) 2419e2426

2421

a 16:1 mass ratio in the 200 g treatments. Representing anticipated bulk availability at Collinsville, sewage consisted of 3:1 mass ratio mix of Bowen to Collinsville sewage. Drying for 24 h at 105  C showed that the organic matter used constituted approximately 48%, 66%, and 90% dry weight of green waste, Bowen sewage and Collinsville sewage respectively. As organic amendments could signicantly increase concentrations of solutes such as nutrients and heavy metals, the maximum contribution each organic material could make was calculated by assuming a worst-case scenario of 100% dissolution of their total mass to the GAE water. 2.4. Sampling and analysis Microcosms were maintained at 25  C, and regularly monitored for a total of 230 days. A Hydrolab Datasonde 4a multiparameter meter (Hydrolab, USA) was used to measure upper microcosm water quality parameters on 10 regular occasions for temperature, Electrical Conductivity (EC), pH, and ORP. A WTW pH meter probe was also inserted down the mesh tube to measure microcosm water pH just above the sedimentewater interface on these same occasions. In order to determine dissolved carbon contribution by waste material to the microcosm water, a 60 mL surface water sample was taken without replacement 8 days after dosing, ltered through 0.5 mm glass bre lter paper (PAL Metrigard) and analysed for dissolved organic carbon (DOC) measured as non-purgeable organic carbon (NPOC) on a total organic carbon (TOC) Analyser (Shimadzu, Japan). A second water sample was taken after 180 days of organic material dosing when water quality remediation appeared to have occurred. This water sample was ltered and then split into two. One sample was frozen prior to analysis for NPOC as stated earlier, sulphate (SO2 4 ) was measured on an ion chromatograph (Model 7961, Metrohm, Switzerland), and for analysis of total lterable nitrogen (TFN) samples were digested with persulfate according to APHA (1998). The second sample was acidied with 1% reagent grade HCl and analysed by Inductively Coupled Plasma Atomic Emission Spectrophotometry (ICP-AES) for Al, As, B, Ca, Cd, Cr, Co, Cu, Fe, Pb, Mg, Hg, S, Se, Si, Sn, Zn. Microcosm sediment geochemistry was sampled on day 230 when the experiment indicated no further pH increase. This was carried out by manually removing surface organic material and homogenising the remaining top 25 mm of sediment of each microcosm core where the bulk of SRB activity products were expected to occur in this experiment time frame (Koschorreck et al., 2007). Sewage, green waste, and microcosm sediment samples were digested as per USEPA (1995). Sediment digestate was analysed by ICP-AES for the same suite of metals as the water samples. Total Kjeldahl N and total P were measured in sewage and green waste samples as per APHA (1998). TOC was measured in sewage and green waste samples on a TOC Analyser 5000a (Schimadzu, Japan). 2.5. Microcosm bacteria and gas sampling Risks which a eld trial of this approach might pose to human health were assessed by measuring total and faecal coliforms and gas production. Day 180 microcosm total coliforms were counted by the standard coliform membrane lter procedure on Sartorius AG 14053 N, endo media (APHA, 1998). Relative SRB abundance was measured on day 180 by pooling 5 mL aliquots of surface water for each treatment in an SRB-BART (Droycon Bioconcepts, Canada) which semi-quantitatively indicated the presence of and measured SRB abundances by a proprietary colourimetric method. Concentrations of the toxic gases of ammonia and hydrogen sulde were measured in the microcosm headspace with Drger tubes (Germany) on days 8, 74 and 125. A 400 mL sample was pooled

together from control replicates, and from separate 200 g treatment replicates (levels of detection to 0.8 ppm). 2.6. Statistical analysis Univariate statistical differences were tested by ANOVA with pair-wise differences determined by Tukeys Honestly Signicant Difference (HSD) post hoc test in SPSS 17 (SPSS Inc., 2009). Multivariate data was analysed by ANalysis-Of-SIMilarity (ANOSIM) and Principal Components Analysis (PCA) ordination were created in PRIMER 6 (PRIMER-E Ltd, 2001). 3. Results 3.1. Organic amendments Sewage sludge was dried by the municipal water treatment plants for 12 and 3 weeks respectively for Collinsville (density 1.25 kg L1) and Bowen (density 1.35 kg L1) sources. Green waste produced a lower paste pH (5.6) than sewage sludge (6.1 and 6.5 for Collinsville and Bowen sources respectively). Sewage contained very high sulphur concentrations (between 10 and 14 g kg1) and was almost an order of magnitude higher than green waste (1.7 g kg1). Total Kjeldahl N was highest in Bowen sewage (60 mg kg1), compared to Collinsville (31 mg kg1) and lowest in green waste (20 mg kg1), with a similar pattern for total P (19, 12 and 2.3 mg kg1 respectively). TOC was highest in green waste at 39%, 34% in Bowen sewage and 29% in Collinsville sewage. Both sewage types were more alkaline than green waste (pH 6.1 and 6.5 versus 5.6 respectively). Metal composition of green waste and Bowen sewage is detailed in McCullough et al. (2006). Table 1 shows organic dosing could potentially make a substantive contribution to nutrient (N and P) availability in the microcosms. Organic dosing could also increase concentrations of As, Ca, and Cr in all treatments. In treatments containing sewage; Al, Cd, Cu, Fe, Pb, Mg, Ni, and Zn were all potentially increased. 3.2. Water chemistry changes Early in the experiment an S200 microcosm core developed a leak rendering only two replicates for this treatment. Temperature was very similar between all treatments, averaging 25.0  C (0.1  C standard error of the mean, S.E.). The pH of the control
Table 1 Maximum possible increase in water solute concentrations (all values mg L1, e No data; * Total nitrogen as Kjeldahl nitrogen). GAE pit lake water, G green waste only, GS green waste and sewage, S sewage only. 100 100 g and 200 200 g loading. GAE Total sulphur Sulphate Total N* Total P DOC Aluminium Arsenic Cadmium Calcium Chromium Cobalt Copper Iron Lead Magnesium Nickel Zinc 2900 2600 0.5 <0.1 5.4 200 0.0 0.0 520 0.1 2.6 0.3 690 0.3 730 5.2 16 G100 24 71 0.3* 0.0 540 14 0.0 0.0 360 0.10 0.2 0.0 9 0.0 39 0.0 0.7 G200 47 140 0.6* 0.1 1100 28 0.0 0.0 720 0.2 0.3 0.0 19 0.0 78 0.0 1.3 GS100 290 870 1.3 0.4 10 230 0.2 0.1 750 0.9 0.2 16 240 1.9 130 0.7 24 GS200 580 1700 3 1 21 470 0 0 1500 2 0 32 49 4 260 1 48 S100 270 800 1.1* 0.4 680 220 0.2 0.1 390 0.8 0.1 16 230 1.9 91 0.7 23 S200 530 1600 2.1* 0.7 1400 440 0.3 0.1 790 1.6 0.2 32 470 3.7 180 1.3 47

2422

C.D. McCullough, M.A. Lund / Journal of Environmental Management 92 (2011) 2419e2426

treatment also changed little over the course of the experiment (0.05 pH units from mean 2.4). Eight days after addition of organic substrates to microcosm cores, microcosm pH values were already highly signicantly different between treatments (F6,17 437, p < 0.001), although not signicantly different between different doses within the same treatment type (Tukeys HSD, p > 0.05) (Fig. 2a). By day 230 end of experiment, pH had not increased signicantly in G treatments at only around pH 4.5. However, aside from S100 which only reached pH 6, in general S and GS treatments showed a rapid increase to circum-neutral pH by day 230. There also appeared to be greater variation in pH within the G treatments than for either GS or S treatments. Although there was generally

little difference between upper and lower pH of microcosms in the rst 30 days after dosing, S200 had higher pH near the sediment and GS100 had higher pH at the water surface. After 60e90 days, the difference in pH through the water column of all treatment microcosms was less than 1 pH unit. EC increased initially in all treatments with highest nal EC in GS200 and S200 at 11.3 and 12.1 mS cm1 respectively (Fig. 2b). Although there was a slight increase in control EC over the experiment from 9.4 to 9.9 mS cm1, EC declined in both G treatments and in S100 at around the same time as pH increased (i.e., around day 37). Although EC variability was very high for S treatments, EC appeared to decline faster in GS and S treatments than in G treatments. Control ORP did not change over the course of the experiment, remaining at 600 mV. However, ORP values of all treatments amended with organic waste declined quickly and were signicantly lower than control just 8 days after addition of organic substrates (F6,17 605, p < 0.001, Tukeys HSD, p < 0.05) (Fig. 2c). Day 8 ORP values were similar for different dosing loads of treatments, except for S200 which had signicantly higher ORP than S100 (Tukeys HSD, p < 0.001). ORP values were low in GS and S treatments containing sewage at between 300 and 400 mV from day 40 to the end of the experiment. This ORP was consistently lower than in G treatments without sewage which were highly variable at around 150 mV. Visually, the controls did not change over the experiment. However, G treatments darkened and produced a sweet odour with a blue-green coloured mould on the water surface. The water in the S treatments also appeared to lose the orange colour found in the controls and an orange precipitate settled on the microcosm walls. After only a few weeks the sediment surface of all treatment microcosms amended with organic waste began to blacken. A PCA of day 180 physico-chemical data indicated that the treatments had improved water quality when compared to the control (Fig. 3). ANOSIM showed water quality was statistically different from control (p < 0.05), except for S200 which had only two replicates remaining. GS showed the greatest water quality improvement followed by S and then G treatments. The PCA showed strong separation of day 180 treatments from controls associated with a change in pH and ORP. With the exception of G treatments, this showed greater organic amendment increased day 180 pH.

Fig. 2. Mean (s.e) change in a) pH, b) EC, and c) ORP of different treatments over time. G green waste only, GS green waste and sewage, S sewage only. 100 100 g and 200 200 g loading. Trend lines tted to aid trend visualisation. Vertical hairline indicates day 8 sampling.

Fig. 3. PCA of day 180 microcosm water quality data, ordination variability explained by each axis shown as a percentage. G green waste only, GS green waste and sewage, S sewage only. 100 100 g and 200 200 g loading. Lines indicate eigenvectors (only those >[0.2] shown), the direction of the line indicates the relationship to a PC axis and the length of the line indicates its relative importance (longer lines more important). One damaged S200 replicate excluded.

C.D. McCullough, M.A. Lund / Journal of Environmental Management 92 (2011) 2419e2426 Table 2 Mean (s.e.) of selected major and minor elements from Day 180 microcosm water (mg L1). G green waste only, GS green waste and sewage, S sewage only. 100 100 g and 200 200 g loading. No s.e. is shown for S200 as only two replicates. * Contains replicates that had concentrations below detection levels e As & Pb < 10 mg L1, Cu & Hg < 1 mg L1, Cd < 0.6 mg L1, Co < 2 mg L1, and Se < 20 mg L1.

2423

Mg K Fe Ca Treatment Al

3.3. Sediment chemistry Day 230 sediment metal concentrations for treatments were less variable than water quality parameters. However, only S treatments showed more variable sediment metal concentrations than control (Table 3). Sediment metal concentrations in the control were only signicantly different (ANOSIM, p < 0.05) to G100 and S treatments. Contrary to the expectation that sulphate reduction would lead to mineral formation in the sediment, GS200 sediments generally contained the lowest concentrations of Al, Ca, Fe, Mg, Mn, Na and S (Table 3). In all other treatments Al, Ca, Mg, and S concentrations increased slightly (110%, 140%, 180% and 130% respectively). Iron concentrations in all treatments were lower than the control, but lowest in GS200. S treatments containing sewage appearing to reduce iron concentrations more than green waste alone. Potassium only decreased in the S-only treatments, to around 80% of control. Microcosm day 230 Fe:S molar ratios decreased 14% in all the treatments compared to controls.

Control G100 G200 GS100 GS200 S100 S200

140 57 66 0.2 0.01 0.02 0.02

9 34 33 0.1 0.01 0.01

400 470 460 370 52 400 395

20 0 20 35 4 11

470 560 280 110 0.5 24 0.2

160 50 230 110 0.1 23

0.8 150 180 170 330 47 82

0.5 30 59 12 20 1

550 580 560 530 440 580 565

19 17 18 26 23 26

The mixture of organics amended in GS substantially increased pH attained by either G or S amendment alone. For example, G100 and S100 produced a similar pH and water quality to S200. Increases in pH were associated with reductions in S and sulphate, ORP, DO, and most metal concentrations, which were indicative of sulphate reduction. The heterogeneous nature of green waste organic amendments was reected in water chemistry variability within G and GS treatments which was higher than S treatments; particularly at the 200 g level. Interestingly, S100 was associated with higher Se, TFN, Mg and Ca concentrations compared to all other treatments. The addition of more organic matter (200 g) to S treatments, resulted in lower EC, despite the addition of more solutes (see Table 1). Day 8 DOC was 2500% and 6000% higher in G100 and G200 (200 40 mg L1 and 480 50 mg L1 mean S.E. respectively) than in control (8.1 2.3 mg L1), 1600%e2500% higher in S100 and S200 (130 15 mg L1 and 200 mg L1 respectively), 4400% higher in GS100 (360 90 mg L1), but just 30% of control in GS200 at 30% in GS200. Substantial improvements in water quality were seen in all treatments by Day 180 compared to control, with GS200 showing the most improvement for all measured major parameters (excluding conservative ions) (Table 2). Total lterable N increased 3000%e110,000% from control in all S treatments, but <4000% in the G and GS200 treatments (Table 2). Furthermore, GS200 also showed strong evidence of sulphate reduction with a seven-fold reduction in sulphate concentrations and reductions to near detection limits for Fe, Zn, Mn, Ca and Al. Sulphate decreased in all treatments, but particularly in 200 g treatments (except G200) with up to 20% concentration reduction in GS200. Aluminium decreased to less than 1% in all S treatments and to 39% and 46% in G100 and G200 respectively. Concentrations of conservative ions of Ca and Mg generally did not signicantly change in treatments, although Ca dropped to around 13% in GS200. Potassium and Na increased in all treatments in proportion to the amount of material added. Fe concentrations decreased most in higher and sewage-containing GS and S treatments to 0.04% in S200 and 0.1% in GS200. Manganese only decreased in GS and S treatments containing sewage, to a minimum of 2% in GS200. Zinc decreased to <50% in the G-only treatments and <2% in all other treatments (Table 2). Cadmium, Co, Cr, Cu, Ni and Pb were highest in control, but concentrations decreased almost 50% in G treatments and were close to detection limits in all S treatments. Boron and Hg increased 280% and 500% respectively in high-dosed treatments. Selenium was similar across all treatments and control. Arsenic decreased 120% in low-dosed treatments and increased to 200% in GS200 and S200 treatments with high levels of sewage.

Cr Cd B
SO2 4

S Cl TFN Zn Na Mn

47 46 43 5.5 0.7 3 3.2

2 1 2 0.5 0.1 3

410 530 470 600 770 580 710

13 13 0.7 0.5 15 4.0 3.5 20 21 6.5 3.4 15 6 0.07 0.01 55 71 0.18 0.06 15 32 0.08 0.04 180 0.05 370

0.2 2 5 48 3 6

63 59 770 830 930 670 890

21 17 44 23 82 6

2100 1600 1600 830 570 1500 1300

71 88 150 49 150 58

9700 4300 7100 3000 1300 5100 4900

940 380 960 170 500 45

96 380 400 320 460 280 330

9 49 58 26 20 17

5.5 2.3 1.5 0.6 0.6 0.6 0.6*

0.3 1.7* 0.9* 0* 0* 0*

37 14 20 2.3 2 1.7 1.5

1 290 34 3700 170 30 0 6 6.7 3.0* 1800 920 20 6* 9 1.3 0.3* 2000 910 30 15* 0.3 10 0* 110 6 10 0* 0 10 3 170 3 10 0* 0.3 1 0* 12 1 10 0* 1* 50 10

Cu

Ni

Pb

2424 Table 3 Mean (s.e.) (mg kg1) of selected major and minor elements from day 230 microcosm sediment. G green waste only, GS green waste and sewage, S sewage only. 100 100 g and 200 200 g loading. * No s.e. is shown for S200 as only two replicates. Cd and Se not shown as below detection limits of 0.02 and 0.7 mg kg1 respectively.

C.D. McCullough, M.A. Lund / Journal of Environmental Management 92 (2011) 2419e2426

0.5 5 2 0.9 0.5 1.1

Mean As, Cr and Pb sediment concentrations varied little (Table 3). Boron increased in all treatments to a maximum of 140% (GS200). Cobalt and Ni increased only in G treatments to maximums of 5% and 4340% compared to control and copper only increased 720% in S treatments. 3.4. Bacteria and gases Faecal coliforms were not observed from any treatments. However, total coliform counts were too abundant to be enumerated in G and GS treatments amended with green waste. Sulphate reducing bacteria activity was only found in treatments amended with sewage (GS100, GS200, S100 and S200). A strong smell of hydrogen sulde also evolved and a black precipitate extended upwards through the mesocosm waters above the green waste and sewage solids, however, there were no detectable traces of ammonia or hydrogen sulde on day 8. However, by day 74 there were 2 ppm NH3 and 12 ppm H2S in the headspace of GS200, and 17 ppm H2S in S200. By day 125 there remained 13 ppm NH3 in GS200, and 7 ppm of NH3 and H2S in S200. These gases were absent in control. 4. Discussion A combination of green waste and sewage successfully remediated low pH AMD water containing high total iron concentrations indicative of strongly iron-buffered acidity (Totsche et al., 2006) to circum-neutral pH. Metal and sulphate concentrations were concomitantly substantially reduced in microcosm AMD water. Although SRB activity may not be as limited by low pH as previously thought (Koschorreck, 2008), rate of pH remediation increased sharply above pH 5. Sulphate concentrations decreased in treatments with organic amendments with concomitant H2S production providing support for biological sulphate reduction as the primary mechanism producing alkalinity and removing sulphate and cations. SRB activity was also inferred by the presence of a black precipitate indicative of iron monosuldes and microcosm sediment Fe and S concentrations approximating the 1.1 molar ratio of microbiallyformed iron monosulde (Kleeberg, 1998; Praharaj and Fortin, 2004). The presence of SRB was observed only in BART tests from organic amended treatments. 4.1. Organic material types In disagreement with other studies, green waste organic amendments showed lower bioremediation rates than sewage sludge, even though the biodegradable fraction of sewage sludge was likely lower than plant material hemicellulose and lignin (Gibert et al., 2004; Harris and Ragusa, 2000). Similarly to this study, other studies have also found combined organic mixtures more effective in AMD bioremediation than single substrates (Neculita et al., 2007). Further research is required to understand why some carbon sources, stimulate sulphate reduction whilst others, of comparable carbon content, do not (Zagury et al., 2006). Sediment total sulphur concentrations did not always increase where water column concentrations of sulphate were reduced. Higher Fe:S ratios in control sediments were likely due to the formation of enriched secondary minerals, principally members of the Jarosite group, which may form under highly acidic conditions (Triantafyllidis and Skarpelis, 2005). Even in the short experimental time frame, at such high mean temperatures Fe:S ratios in treatments are likely to have been due to the formation of depleted iron-sulphur compounds such as amorphous iron monosulde (FeS). However, although the intention of removing gelatinous material from the sediment surface was to prevent confounding by the remaining organic matter,

Zn Ni Cu Co Molar ratio Fe:S S Na Mn Mg K Fe Ca Treatment Al B

Control G100 G200 GS100 GS200 S100 S200*

2000 2100 2100 1700 1700 2000 1900

58 180 180 25 31 73 79

1700 1900 1800 2300 1400 1900 1900

204 340 52 270 30 250 230

9100 8800 8500 7900 7000 7100 7200

650 1600 2600 900 600 420 130

1100 1100 1200 1000 1100 940 860

23 110 130 26 62 18 91

600 1000 770 800 560 1100 820

80 74 140 52 43 250 140

65 95 68 44 40 100 48

8 8 8 4 4 9 8

680 630 550 770 610 920 690

700 180 150 64 56 24 320

5000 6300 5700 6100 4700 5000 6000

620 850 1600 240 290 430 660

1.1 0.8 0.8 0.7 0.8 0.8 0.8

0.1 0.2 0.1 0.1 0.1 0.0

5.9 6.9 7.2 8.1 8.5 7.9 7.3

0.7 2.0 0.8 1.2 0.1 1.1

2.5 12 7.3 2.6 1.8 2.9 2.5

0.0 0.0 0.0 0.0 0.0 0.0

2.9 2.7 2.6 6.0 11 17 21

0.4 0.2 0.2 2.0 4 7

4.6 20 12 4.2 2.9 6.2 4.1

37 50 72 44 48 79 44

2.5 25 13 9.0 2.4 21

C.D. McCullough, M.A. Lund / Journal of Environmental Management 92 (2011) 2419e2426

2425

it appears that much of the amorphous metal monosuldes expected to account for the removal of sulphate and metal solutes were associated with this organic material and were also removed, confounding results. For this reason, sediment mineralogy was not further investigated by analyses such as XRD or sequential extractions. This study found organic amendment as sewage sludge alone performed similarly to that of a combination of green waste and sewage, whilst others have reported that vegetation alone preformed equally well (McCullough et al., 2006; Waybrant et al., 1998). This may be due to the fresh nature of the sewage used in this study, with a consequent high labile carbon component that had not yet been mineralised by anaerobic bacterial activity and also with the sewage acting as an inoculum for SRB (Sahinkaya, 2009). Nevertheless, remediation strategies using sulphate reduction ecotechnologies may be best to choose green waste as the bulk contribution to electron donors over sewage as green waste also has an additional advantage of providing complex organic substances and ligand formation between these and heavy metal solutes may further remove these toxic components from biological availability (McCullough, 2007). The use of more refractory organic materials such as green waste may be advantageous to long-term remediation efforts as they will continue to slowly degrade into more labile fractions available for sulphate reducers over longer periods of time (Gibert et al., 2004). 4.2. Health concerns and treatment issues Addition of fresh sewage to acid pit lakes, with or without complementary additions of green waste, may prove to be a novel and practicable remediation strategy for AMD issues in warm-climate mining locations. Nevertheless, there are many environmental and health considerations for the use of sewage in remediating pit lakes. Depending on its source, sewage sludge can contain high concentrations of heavy metals (Berrow and Webber, 1971). These metal loads may prove legally challenging for mining companies which may have to accept responsibility for hazardous materials to be deposited onsite, as part of the remediation process. However, high heavy metal concentrations are unlikely to remain biologically available after bioremediation. Conversion of heavy metals to suldic minerals will prevent remobilisation if anoxic sediments are not re-oxidised (Frmmichen et al., 2004; Wendt-Potthoff et al., 2002). Although release of ammonia gas from the treatment is unlikely to be signicant, release of toxic hydrogen sulde may occur when dissolved concentrations of iron are signicantly decreased. Another concern of use of sewage is the potential for workers to be exposed to human pathogens. Townsville, only 170 km to the northenorth-west of Collinsville, receives a very high mean annual ultraviolet (UV) exposure of 4300 J m2, and a maximum daily exposure of around 60 J m2 during December (personal communication John Javorniczky, Australian Radiation Protection and Nuclear Safety Agency). These UV levels are indicative of the high UV levels likely to be experienced in tropical pit lake surface waters. Consequently, deactivation of pathogens in the warm saline waters of the pit lakes under this high level of UV exposure is likely to be very rapid (Sinton et al., 1999), occurring within days. Furthermore, even if bacteria remain unexposed in sediments and deeper waters, the mesocosm results (conducted under minimal UV exposure) indicate that faecal bacteria are likely to become sublethally injured or die following extended periods in AMD waters (DePasquale et al., 1987; Double and Bissonette, 1980; Winfrey et al., 2010). All treatment faecal coliform abundances were also below national primary contact standards (ANZECC/ARMCANZ, 2000). Long-term potential problems with eld application of sulphate reduction ecotechnology as an in situ treatment of AMD contaminated

pit lakes include ongoing lake acidication through inputs of acidity from the surface and groundwater, and/or through establishment of an acidity driven iron cycle (Geller et al., 2009). This acidity cycle is dependent on Fe2 resulting from anaerobic bacterial reduction of Fe3 in the absence of Fe precipitation and limited SRB activity at low pH. Upon realising an oxidising environment, the mineral Schwertmannite is then formed. In this model, the benets of addition of organic matter to the sediment would eventually be overcome as the cycle became established. Consequently, depth and duration of pit lake water column stratication will remain an important determinant in the efcacy and viability of sulphate reduction as a remediation measure in any pit lake. Although the pit lake epilimnion may deepen under strong forcing winds, the large mass of woody green waste resting on the benthos in a eld-scale trial (Geller et al., 2009), coupled with a high biochemical oxygen demand remaining from ongoing organic decomposition and ongoing carbon diagenesis may nevertheless also provide for hypolimnetic stability. 4.3. Conclusions The high performance of the sewage when combined with green waste suggests that the sewage likely acted as a good inoculum of SRBs and also had carbon in a highly labile form. Day 8 DOC concentrations were highest in all the green waste-treatments organic amendments probably from leached chlorophyll and sap sugars providing a short-lived labile carbon source for initiation of sulphate reduction (Neculita et al., 2007). Refractory green waste may act as a physical substrate for bacterial attachment, although other researchers have not found this to be the case (Koschorreck et al., 2002) as well as providing sustained release of labile carbon as its lignin and hemi-/cellulose contrents slowly degrade. Issues with hydrogen sulde gas formation could be feasibly controlled by introducing low pH iron-laden AMD water to react with excess hydrogen sulde and low water pH to that sub-optimal for SRB activity. Similarly, high abundances of total coliforms recorded in green waste-only samples indicated they were non-pathogenic decomposers that occur naturally in benthic sediments of many natural water bodies. Nevertheless, the extremely high total coliform count (greater than could be enumerated) may mean that further organic decomposition over 180 days is required before primary i.e., human skin contact could be recommended in pit lakes remediated with green waste. Acknowledgements Thanks to Collinsville Coal Project (Xstrata Coal Queensland), particularly Joel May for supporting the project. Thanks also to Joseph Steenbergen and Carlieke te Beest for their assistance in the laboratory with this project. This work was improved with critical reading by Dr. Naresh Kumar and anonymous reviewers. This work was funded by Australian Coal Association Research Program (ACARP) grant C14052 and supported by the Centre of Excellence for Sustainable Mine Lakes, a Western Australian State Government initiative. References
ANZECC/ARMCANZ, 2000. Australian and New Zealand Guidelines for Fresh and Marine Water Quality. National Water Quality Management Strategy Paper No 4. Australian and New Zealand Environment and Conservation Council & Agriculture and Resource Management Council of Australia and New Zealand, Canberra, pp. 1500. APHA, 1998. Standard Methods for the Examination of Water and Wastewater. American Water Works Association, Water Environment Federation, twentieth ed. American Public Health Association, Washington DC, USA. Berrow, M.L., Webber, J., 1971. Trace elements in sewage sludges. J. Sci. Food Agric. 23, 93e100.

2426

C.D. McCullough, M.A. Lund / Journal of Environmental Management 92 (2011) 2419e2426 McCullough, C.D., 2007. Approaches to remediation of acid mine drainage water in pit lakes. Int. J. Min. Reclam. Environ. 22, 105e119. McCullough, C.D., Hunt, D., Evans, L.H., 2009. Sustainable development of open pit mines: creating benecial end uses for pit lakes. In: Castendyk, D., Eary, T., Park, B. (Eds.), Mine Pit Lakes: Characteristics, Predictive Modeling, and Sustainability Society for Mining Engineering (SME). Kentucky, USA, pp. 249e268. McCullough, C.D., Lund, M.A., 2006. Opportunities for sustainable mining pit lakes in Australia. Mine Water Environ. 25, 220e226. McCullough, C.D., Lund, M.A., May, J.M., 2006. Microcosm testing of municipal sewage and green waste for full-scale remediation of an acid coal pit lake, in semi-arid tropical Australia. In: Barnhisel, R.I. (Ed.), Proceedings of the 7th International Conference on Acid Rock Drainage (ICARD). American Society of Mining and Reclamation (ASMR), St Louis, Missouri, USA, pp. 1177e171, 197. McCullough, C.D., Lund, M.A., May, J.M., 2008a. Field scale demonstration of the potential for sewage to remediate acidic mine waters. Mine Water Environ. 27, 31e39. McCullough, C.D., Lund, M.A., May, J.M., 2008b. Field scale trials treating acidity in coal pit lakes using sewage and green waste. In: Rapantova, N., Hrkal, Z. (Eds.), Proceedings of the 10th International Mine Water Association (IMWA) Congress. Karlovy Vary, Czech Republic, pp. 599e602. Neculita, C.-M., Zagury, G.J., Bussire, B., 2007. Passive treatment of acid mine drainage in bioreactors using sulfate-reducing bacteria: critical review and research needs. J. Environ. Qual. 36, 1e16. Praharaj, T., Fortin, D., 2004. Indicators of microbial sulfate reduction in acidic sulde-rich mine tailings. Geomicrobiol. J. 21, 457e467. PRIMER-E Ltd, 2001. PRIMER, 5.0 ed. Plymouth Marine Laboratory, Plymouth. Sahinkaya, E., 2009. Microbial sulfate reduction at low (8  C) temperature using wastesludge as a carbon and seed source. Int. Biodeter, Biodegrad. 63, 245e251. Sinton, L.W., Finlay, R.K., Lynch, P.A., 1999. Sunlight inactivation of fecal bacteriophages and bacteria in sewage-polluted seawater. Appl. Environ. Microbiol. 65, 3605e3613. SPSS Inc., 2009. SPSS, 17.0 ed. SPSS Inc., Chicago. Totsche, O., Fyson, A., Steinberg, C.E.W., 2006. Microbial alkalinity production to prevent reacidication of neutralized mining lakes. Mine Water Environ. 25, 204e213. Triantafyllidis, S., Skarpelis, N., 2005. Mineral formation in an acid pit lake from a high-suldation ore deposit: Kirki, NE Greece. J. Geochem. Explor. 88, 68e71. USEPA, 1995. SW-846 EPA Method 3050B: Acid Digestion of Sediments, Sludges, and Soils, Test Methods for Evaluating Solid Waste, third ed. USEPA, Washington, DC. 3rd update. Waybrant, K.R., Blowes, D.W., Ptacek, C.J., 1998. Selection of reactive mixtures for use in permeable reactive walls for treatment of mine drainage. Environ. Sci. Technol. 32, 1972e1979. Wendt-Potthoff, K., Frmmichen, R., Herzsprung, P., Koschorreck, M., 2002. Microbial Fe(III) reduction in acidic mining lake sediments after addition of an organic substrate and lime. Water Air Soil Pollut. Focus 2, 81e96. Wielinga, B.W., 2009. In situ bioremediation of pit lakes. In: Castendyk, D., Eary, T. (Eds.), Handbook of Technologies for Management of Metal Mine and Metallurgical Process Drainage. Mine Pit Lakes: Characteristics, Predictive Modeling, and Sustainability, vol. 3. Society for Mining, Metallurgy, and Exploration, Colorado, USA, p. 412. Winfrey, B.K., Strosnider, W.H., Nairn, R.W., Strevett, K.A., 2010. Highly effective reduction of fecal indicator bacteria counts in an ecologically engineered municipal wastewater and acid mine drainage passive co-treatment system. Ecol. Eng. 36, 1620e1626. Zagury, G.J., Kulnieks, V., Neculita, C.-M., 2006. Characterisation and reactivity assessment of organic substrates for sulphate-reducing bacteria in acid mine drainage treatment. Chemosphere 64, 944e954.

Castro, J.M., Moore, J.N., 1997. Pit lakes: their characteristics and the potential for their remediation. Environ. Geol. 39, 254e260. Commonwealth of Australia Bureau of Meteorology, 09/02/2009. Collinsville Climate Averages. DePasquale, D.A., Law, C.B., Bissonette, G.K., 1987. Comparative survival and injury of Candida albicans and bacterial indicator organisms in streams receiving acid mine drainage. Water Res. 21, 1525e1530. Double, M.L., Bissonette, G.K., 1980. Enumeration of coliforms from streams containing acid mine water. Res. J. Water Pollut. C 52, 1947e1952. Doup, R.G., Lymbery, A.J., 2005. Environmental risks associated with benecial end uses of mine lakes in southwestern Australia. Mine Water Environ. 24, 134e138. Drury, W.J., 1999. Treatment of acid mine drainage with anaerobic solid substrate reactors. Water Environ. Res. 71, 1244e1250. Frmmichen, R., Wendt-Potthoff, K., Friese, K., Fischer, R., 2004. Microcosm studies for neutralization of hypolimnic acid mine lake water (pH 2.6). Environ. Sci. Technol. 38, 1877e1887. Geller, W., Koschorreck, M., Wendt-Potthoff, K., Bozau, E., Herzsprung, P., Bttner, O., Schultze, M., 2009. A pilot-scale eld experiment for the microbial neutralization of a holomictic acidic pit lake. J. Geochem. Explor. 100, 153e159. Gibert, O., de Pablo, J., Cortina, J.L., Ayora, C., 2002. Treatment of acid mine drainage by sulphate-reducing bacteria using permeable reactive barriers: a review from laboratory to full-scale experiments. Rev. Environ. Sci. Biotechnol. 1, 327e333. Gibert, O., de Pablo, J., Cortina, J.L., Ayora, C., 2004. Chemical characterisation of natural organic substrates for biological mitigation of acid mine drainage. Water Res. 38, 4186e4196. Gray, N.F., 1997. Environmental impact and remediation of acid mine drainage: a management problem. Environ. Geol. 30, 62e71. Harris, M.A., Ragusa, S., 2000. Bacterial mitigation of pollutants in acid drainage using decomposable plant material and sludge. Environ. Geol. 40, 195e215. Hobbs, P., Oelofse, S.H.H., Rascher, J., 2008. Management of environmental impacts from coal mining in the Upper Olifants River catchment as a function of age and scale. Int. J. Water Resour. Dev. 24, 417e431. Klapper, H., Geller, W., 2002. Water quality management of mining lakes e a new eld of applied hydrobiology. Acta Hydroch. Hydrob. 29, 363e374. Kleeberg, A., 1998. The quantication of sulfate reduction in sulfate-rich freshwater lakes e a means for predicting the eutrophication process of acidic mining lakes. Water Air Soil Pollut. 108, 365e374. Kolmert, ., Johnson, D.B., 2001. Remediation of acidic waste waters using immobilised, acidophilic sulfate-reducing bacteria. J. Chem. Technol. Biotechnol. 76, 836e843. Koschorreck, M., 2008. Microbial sulphate reduction at a low pH. FEMS Microbiol. Ecol. 64, 329e342. Koschorreck, M., Bozau, E., Frommichen, R., Geller, W., Herzsprung, P., WendtPotthoff, K., 2007. Processes at the sediment water interface after addition of organic matter and lime to an acid mine pit lake mesocosm. Environ. Sci. Technol. 41, 1608e1614. Koschorreck, M., Frmmichen, R., Herzsprung, P., Tittel, H., Wendt-Potthoff, K., 2002. The function of straw for in situ remediation of acidic mining lakes: results from an enclosure experiment. Water Air Soil Pollut. 2, 97e109. Kumar, R.N., McCullough, C.D., Lund, M.A. How does storage affect the quality and quantity of organic carbon in sewage for use in the bioremediation of acidic mine waters? Ecol. Eng., in press-a. doi:10.1016/j.ecoleng.2011.02.021. Kumar, N.R., McCullough, C.D., Lund, M.A., Newport, M. Sourcing organic materials for pit lake remediation in remote mining regions. Mine Water Environ., in press-b. doi:10.1007/s10230-011-0144-6. Ksel, K., Dorsch, T., 2000. Effect of supplemental electron donors on the microbial reduction of Fe(III), sulfate, and CO2 in miningeimpacted freshwater lake sediments. Microb. Ecol. 40, 238e249. Martins, M., Faleiro, M., Barros, R., Verssimo, A., Costa, M., 2009. Biological sulphate reduction using food industry wastes as carbon sources. Int. Biodeter. Biodegrad. 20, 559e567.