Environ. Sci. Technol. 2007, 41, 2407-2413
Arctic Vegetation Damage by Winter-Generated Coal Mining Pollution Released upon Thawing B O E L B E R L I N G , * ,† J E N S S Ø N D E R G A A R D , † LOUISE A. JENSEN,† LEA B. SCHMIDT,† BIRGER U. HANSEN,† GERT ASMUND,‡ T O N C I B A L I CÄ - Z U N I CÄ , § JØRGEN HOLLESEN,† SUSANNE HANSON,† PER-ERIK JANSSON,| AND THOMAS FRIBORG† Institute of Geography and Geology, University of Copenhagen, Copenhagen, Denmark, Department of Arctic Environment, National Environmental Research Institute, Roskilde, Denmark, and Department of Land and Water Resources Engineering, Royal Institute of Technology, Stockholm, Sweden
Acid mine drainage (known as AMD) is a well-known environmental problem resulting from the oxidation of sulfidic mine waste. In cold regions, AMD is often considered limited by low temperatures most of the year and observed environmental impact is related to pollution generated during the warm summer period. Here we show that heat generation within an oxidizing, sulfidic, coal-mining wasterock pile in Svalbard (78° N) is high enough to keep the pile warm (roughly 5 °C throughout the year) despite mean annual air temperatures below -5 °C. Consequently, weathering processes continue year-round within the wasterock pile. During the winter, weathering products accumulate within the pile because of a frozen outer layer on the pile and are released as a flush within 2 weeks of soil thawing in the spring. Consequently, spring runoff water contains elevated concentrations of metals. Several of these metals are taken up and accumulated in plants where they reach phytotoxic levels, including aluminum and manganese. Laboratory experiments document that uptake of Al and Mn in native plant species is highly correlated with dissolved concentrations. Therefore, future remedial actions to control the adverse environmental impacts of cold region coal-mining need to pay more attention to winter processes including AMD generation and accumulation of weathering products.
Introduction Coal mining has been an important industry in the Arctic for more than 100 years and is likely to expand due to global needs and the abundance of coal in this region (1-3). The environmental impacts from coal mining are well-known and include soil acidity, toxic metal concentrations, and vegetation damage (4-7). The cold climate and permafrost * Corresponding author address: Øster Voldgade 10, DK-1350 Copenhagen K, Denmark; phone: +45 353 22520; fax: +45 3532 2501; e-mail:
[email protected]. † Institute of Geography, University of Copenhagen. ‡ National Environmental Research Institute. § Institute of Geology, University of Copenhagen. | Royal Institute of Technology. 10.1021/es061457x CCC: $37.00 Published on Web 02/24/2007
2007 American Chemical Society
in the Arctic are often considered as factors that may reduce element cycling (8-10) and the release of pollutants throughout most of the year (11). Environmental problems associated with coal mine waste disposal are related to oxidation of pyrite (FeS2) which produces sulfuric acid and triggers a release of heavy metals. Such acid sulfate- and metal-rich solutions are known as acid mine drainage, AMD (11, 12). Depending on the rate of pyrite oxidation and the availability of water and oxidants, as well as buffering capacity of surrounding sediments, the oxidation of sulfide minerals and the subsequent acid production may lead to generation of acidic leachates containing high concentrations of Fe, SO4, Al, and heavy metals, including Pb, Zn, Cd, etc. (13, 14). The discharge of acid drainage into rivers, lakes, and marine waters may cause an instant threat to the biota and the ecological balance by a number of direct and indirect pathways (15-17). Therefore, the disposal of mine tailings in an environmentally sound yet cost-effective manner, aiming to control AMD generation and to limit the environmental impact of AMD, is an important challenge for the mining industry. Several options are currently being examined for the abatement and treatment of AMD. Some methods are directed toward the treatment of the resulting drainage, while others are directed toward prevention of AMD generation. In cold regions, methods include the use of inert covering (18), freezing (19), and subaqueous disposal (20, 21). However, attempts to prevent AMD generation in the Arctic have been made without an understanding of environmental controls of seasonal trends; in particular, controls functioning during the winter are poorly understood. We hypothesize that current predictions of winter and total pollution from wasterock piles in cold regions are underestimated and that wintergenerated pollution, sustained by heating due to pyrite oxidation, in many cases is more harmful to the environment when released over a short period of time during initial thawing than pollution generated during the summer. To test the hypothesis we have studied the production of acid mine drainage and its impact on vegetation at an abandoned coal mining waste-rock pile in Bjørndalen near Longyearbyen in Svalbard (78° N).
Experimental Section Study Site and Sampling. More than 40 similar rock piles are found in close proximity to the pile investigated in Bjørndalen near Longyearbyen (78° N). The rock pile studied is 20 m high and consists of roughly 200 000 m3 of waste rocks deposited between 1986 and 1996. Volume- and depthspecific samples of the pile were collected with a 0.5 m increment from one central hole to a depth of 7 m and two replicate holes to a depth of 3 m in August 2004. Subsamples (finer than 2 mm but accounting for more than 98% of the total specific surface area) were brought to Denmark for analysis. The central site was instrumented to measure gas composition, water content, and temperature every 0.5 m to a depth of 7 m. A weather station was placed on the middle of the pile and has logged subsurface water contents and temperatures, as well as air temperature, radiation, wind, humidity, snow thickness, and atmospheric pressure every half hour. Downstream from the waste-rock pile, 24 sites were selected along two transects in August-September 2004. One transect runs parallel to the elongation of the pile, assessing the degree of pollution from unaffected to affected sites. A second transect runs perpendicular to the elongation of the pile to assess trends with increasing distance from the pile (Supporting Information, Figure 1). At each site the aboveVOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2407
ground plant biomass was quantified on a species-specific basis and harvested. Volume- and depth-specific soil samples were collected to a depth of 50 cm, and the fraction finer than 2 mm was further analyzed. At the end of October 2004, a trench was made below the pile to collect the runoff during the 2005 season and channel it 20 m downstream to a constructed weir, where water discharge was logged and measured manually. Water samples were taken, and in situ measurements of pH and conductivity were made daily over the entire runoff period. Biogeochemical and Mineralogical Analyses of Waste Material. Subsamples from the three profiles within the waste-rock pile (26 samples) were analyzed for pH at a ratio of 20 g shaken in 50 mL of water (1:2.5) and for mineral composition using X-ray powder diffraction (Philips PW3710). Minerals were identified by comparison of powder diffraction patterns with calculated patterns based on the structural data in the Inorganic Crystal Structure Database (ICSD, produced by Gmelin-Institut fu ¨ r Anorganische Chemie and Fachinformationszentrum FIZ, Karlsruhe, Germany). Samples were also analyzed for major elements as well as trace elements by X-ray fluorescence with a PANalytical MagiX PRO spectrometer according to ref 22. The potential acid/ base generation was evaluated by the classical acid-base accounting overburden analysis (23). Pyrite oxidation rates were measured calorimetrically (24) under aerobic conditions with a thermal-activity monitor (type 2277, Thermometric, Sweden, or C3-analysentechnik, Germany) equipped with ampule cylinders (4 mL twin, type 2277-201, and 20 mL twin, type 2230). Glass ampoules containing 10 g of waste in air were inserted in the measurement cylinders, and after thermal equilibration, the heat output was recorded before and after microbes were exterminated with chloroform. Biotic rates were calculated as the difference between total and abiotic rates. The presence of pyrite-oxidizing microbes was assessed by counting the number of Acidithiobacillus ferrooxidans cells (per gram dry weight) on incubated agar plates (24). Element Analysis of Soil, Plants, and Runoff Water. Element analyses of eight major and trace constituents (Al, Fe, Mn, Zn, Ni, Cr, As, and Pb) were made on soil, plants, and runoff water by inductively coupled plasma mass spectrometry (ICP-MS; Agilent 7500CS). Additional analyses of Cu, Cd, K, Na, Mg, and Ca were made on runoff water. Analysis of soil samples were made after the extraction of the bioavailable fraction with 25% acetic acid as suggested (25) as a measure of ion-exchangeable ions, which may be more or less bound to amorphous compounds of iron and manganese and carbonates. Total concentrations of elements in soil samples were extracted with hydrofluoric acid. Siteand species-specific vegetation samples were washed thoroughly with deionized water, dried, and divided into above-ground and below-ground fractions before they were blended and mixed with Superpure HNO3 and thereafter microwave-digested. Filtered runoff water was acid-treated prior to measurements. Concentrations of anions were analyzed on non-acid-treated samples by Ion chromatography (Metrohm 761 Compact IC), and total dissolved carbon was determined with a Shimadzu 5000A total organic carbon analyzer. Heat Generation and Subsurface Temperature Modeling. The COUP model is a well-documented one-dimensional numerical model (26) that previously has been used successfully to predict subsurface temperatures in coarse material deposited in Arctic soils (19). For the purpose of this work, the model was used to simulate subsurface temperatures based on weather conditions at the surface, include air temperature, precipitation, energy balance, and relative humidity as well as subsurface spatial distribution of heat generation, grain-size distribution, and water content. Simulations were made by dividing the waste-rock pile into 2408
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 7, 2007
22 layers; each layer was characterized with respect to hydraulic and thermal properties. The model was successfully calibrated and validated consistent with similar simulations (19) and based on correlation analysis of simulated and measured subsurface temperatures from October 2004 to March 2006. The first winter period was used for calibration and the second winter period for validation. Simulations were made from 1986 to 2006,with the assumption that the initial waste-rock pile temperature was 0 °C. Input data for the period 1986-2004 were based on daily climate data from the Longyearbyen Airport located 5 km away (www.eklima.no). Little or no snow accumulates on the pile because it is usually blown away. Thus, simulated subsurface temperatures are the best conservative estimate for observations in the study year 2004-2005. However, actual subsurface temperatures are likely to be even higher in years where a snow cover is established. Laboratory Vegetation Experiment. Experiments were made in a Termaks KBP 6395L climatic growth chamber at 10 °C, with constant humidity, and with plants exposed to photosynthetic available radiation (approximately 400 µmol m-2 s-1) for 24 h per day. Experiments included two graminoid species in 30 replicates: Poa arctica, which was found in unpolluted sites only, and Luzula confusa, which was observed in unpolluted and polluted sites. Graminoids of equal size used for the experiments were removed from an unpolluted site and replanted in plastic flowerpots with unaffected reference soil found in Bjørndalen. Flower pots were watered with unpolluted water for 2 weeks and subsequently watered every fourth day with 25 mL of either (1) unpolluted water (from Isdammen, the drinking water resource for Longyearbyen); (2) 100% polluted water collected downstream from the pile, representing a snapshot of element concentrations in runoff; or (3) 50% polluted water, made as a 1:1 mix of the other two (Table 1). Graminoids were harvested after 7 months of treatment, and the dried aboveground biomass was analyzed for metal contents. Speciation Modeling. The presence of free ions was evaluated by PHREEQC (27), which calculates the geochemical equilibrium based on thermodynamic data given for temperatures between 1 and 100 °C. Inputs for simulations include total concentrations of dissolved elements as well as pH, redox pe, and temperature. Cation-anion balance of the runoff water was confirmed (2 mm; consisting of calcite and dolomite). Secondary minerals include jarosite, rozenite, and gypsum, which indicate fairly dry conditions. The waste-rock pile consists of rocks more than 0.5 m in diameter but is dominated by 1-10 mm rocks (Supporting Information, Figure 2). Element analysis of samples collected with a 0.5 m increment could not reveal any significant trends with depth or between the three profiles (Supporting Information, Figures 3 and 4). Major element concentrations (percentage ( 1 standard deviation) were on average 0.53 ( 0.08 Na2O, 0.60 ( 0.39 MgO, 7.24 ( 1.89 Al2O3, 30.28 ( 7.83 SiO2, 4.35 ( 2.10 S, 1.49 ( 0.82 CaO, 0.01 ( 0.02 MnO, and 7.09 ( 2.73
TABLE 1. Geochemical Characteristics Downstream from the Coal Mining Pilein Summer 2004 environmental impact areas low impact (n ) 9) mean pH
7.6
Al Cr Mn Fe Ni Zn As Pb
2462.5 5.6 128.1 7238.6 5.1 29.3 0.3 2.0
Al Cr Mn Fe Ni Zn As Pb
1201.5 1.6 121.5 925.8 7.1 41.7 1.0 0.5
woody plants graminoids mosses bare soil lichens
5.0 33.7 25.1 34.7 1.4
unpolluted (µg L-1) pH Al Cr Mn Fe Ni Zn As Pb a
6.2 0.5 0.3 0.2 47.1 0.4 2.9 0.0 0.0
medium impact (n ) 7)
std dev
mean
0.6
pH A-horizon 5.8
high impact (n ) 7)
std dev
mean
std dev
0.8
3.0
0.4
A-horizon Bioavailable Concentrations (mg kg-1) 801.8 3272.5 1285.6 1.3 6.8 2.4 51.4 307.6 265.7 2567.6 13 032.1 3500.0 2.6 63.7 77.6 13.8 43.9 25.5 0.1 1.6 2.4 0.8 1.4 0.5
2617.0 6.9 73.7 16 688.3 5.2 39.7 0.6 0.5
1086.9 1.7 33.0 2779.6 3.1 15.2 0.3 0.5
1122.5 1.5 57.1 731.7 1.7 24.5 0.5 0.5
1922.4 1.8 373.2 2315.2 27.9 163.0 5.9 0.7
397.9 0.6 187.0 1404.0 8.1 73.1 5.4 0.2
0.0 0.0 19.5 33.1 40.1
0.0 0.0 0.0 59.0 41.0
0.0 0.0 0.0 40.4 40.4
Plant Concentrationsa (mg kg-1) 658.0 1201.6 0.9 1.3 83.2 191.4 656.7 738.7 8.3 5.7 27.6 61.2 0.8 0.6 0.3 0.6 5.1 33.2 17.0 30.3 3.8
Coverage (%) 0.0 0.0 15.3 58.3 26.3
Experimental Water Treatment 50% (mg L-1) 3.5 119.3 0.0 21.1 1.3 2.2 7.2 0.5 (µg L-1) 1.0 (µg L-1)
100% (mg L-1) 3.2 253.0 0.0 46.3 4.3 4.5 13.6 1.0 (µg L-1) 2.0 (µg L-1)
Average concentration in Poa arctica, Luzula confusa, and Phippsia concinna.
Fe2O3. A significant correlation (R2 ) 0.82, p < 0.05) was noted between Si and Al (mainly silicate minerals), whereas no significant correlation was noted between other elements. Metal concentrations [parts per million (ppm) ( 1 standard deviation] were on average 18 ( 5 Pb, 23 ( 14 Zn, 10 ( 22 Cu, 13 ( 9 Ni, and 40 ( 17 Mn. Calorimetric results indicate an average heat generation rate of the bulk waste of 0.04 ( 0.02 W m-3 (at 5 °C) and a temperature sensitivity (i.e., Q10) of 2.6, in accordance with other investigations (24). About 60% of the heat was related to microbial activity, which is consistent with the presence of 4000-36 000 Acidithiobacillus ferrooxidans (number of cells g-1). Since September 2004 the water content has remained low (less than 18% by volume) and pore air O2 gas concentrations stayed above 19% (Supporting Information, Figure 5). The average acid production potential was 144 ( 30 mequiv of H+ (100 g of sample)-1. With no measurable neutralization potential of the fine fraction, the waste is net acid producing which is in line with low measured pH values of waste mixed with water (
