Applied Geochemistry (v.25, #3)

Abundance and fractionation of Al, Fe and trace metals following tidal inundation of a tropical acid sulfate soil by Scott G. Johnston; Edward D. Burton; Richard T. Bush; Annabelle F. Keene; Leigh A. Sullivan; Douglas Smith; Angus E. McElnea; Col R. Ahern; Bernard Powell (323-335).
Tidal inundation was restored to a severely degraded tropical acid sulfate soil landscape and subsequent changes in the abundance and fractionation of Al, Fe and selected trace metals were investigated. After 5 a of regular tidal inundation there were large decreases in water-soluble and exchangeable Al fractions within former sulfuric horizons. This was strongly associated with decreased soil acidity and increases in pH, suggesting pH-dependent immobilisation of Al via precipitation as poorly soluble phases. The water-soluble fractions of Fe, Zn, Ni and Mn also decreased. However, there was substantial enrichment (2–5×) of the reactive Fe fraction (FeR; 1 M HCl extractable) near the soil surface, plus a closely corresponding enrichment of 1 M HCl extractable Cr, Zn, Ni and Mn. Surficial accumulations of Fe(III) minerals in the inter-tidal zone were poorly crystalline (up to 38% FeR) and comprised mainly of schwertmannite (Fe8O8(OH)6SO4) with minor quantities of goethite (α-FeOOH) and lepidocrocite (γ-FeOOH). These Fe (III) mineral accumulations provide an effective substrate for the adsorption/co-precipitation and accumulation of trace metals. Arsenic displayed contrary behaviour to trace metals with peak concentrations (∼60 μg g−1) near the redox minima. Changes in the abundance and fractionation of the various metals can be primarily explained by the shift in the geochemical regime from oxic–acidic to reducing-circumneutral conditions, combined with the enrichment of reactive Fe near the soil surface. Whilst increasing sequestration of trace metals via sulfidisation is likely to occur over the long-term, the current abundance of reactive Fe near the sediment–water interface favours a dynamic environment with respect to metals in the tidally inundated areas.

Remediation of mine impacted water (MIW) generally requires decreasing the acidity and concentrations of dissolved and/or particulate contaminants ( SO 4 2 - , metals and metalloids). By fulfilling these requirements in both laboratory and field trials, the sustainable composite waste material, crab-shell chitin complex (CC) has proven to be a promising substrate for MIW remediation, but has not yet been directly compared with other substrates under controlled conditions. In this study, remediation rates and metal removal mechanisms promoted by CC were evaluated and compared to the more commonly used lactate and spent mushroom compost (SMC) using sacrificial batch microcosms and geochemical modeling. Under comparable conditions with equivalent mass of substrate to water ratios, increases in pH were much faster in the microcosms containing CC than with the other substrates: CC increased the pH from pH 3.0 to near neutral in 3 d. In microcosms containing CC, steady alkalinity generation and acidity removal were observed at average rates of 26.5 and −25.2 mg CaCO3/L-d, respectively. The activity of SO 4 2 - -reducing bacteria was evident after 9 d of incubation, with average reduction rates of −17.8 mg SO 4 2 - /L-d. Similar changes in alkalinity, acidity, and SO 4 2 - were also observed in lactate-containing microcosms, but only after a 27 d lag period. No alkalinity generation or SO 4 2 - reduction activity was observed in bottles containing SMC. Aluminum removal (100%) was eventually observed with all substrates, but occurred much faster with CC. Results from thermodynamic geochemical modeling indicate that Al removal was consistent with the precipitation of hydroxides and/or alunite. Iron removal was consistent with precipitation of Fe(III) oxides and Fe(II) sulfides, as well as sorption onto CC and SMC. The addition of Na lactate interfered with such mechanisms due to complexation effects. Chitin complex was the only substrate able to partially remove Mn (>73%), likely due to the formation of rhodochrosite. The results of this study indicate that CC is an attractive substrate for treating metal-laden waste streams, especially those which are high in Mn.

To investigate the possible variations of Rn concentration in crystalline rocks as a function of flow conditions, a field study was carried out of a fractured aquifer in granite. The method is based on the in situ measurement of Rn in groundwater, aquifer tests for the determination of hydraulic characteristics of the aquifer and laboratory measurement of Rn exhalation rate from rocks. A simple crack model that simulates the Rn concentration in waters circulating in a fracture intersecting a borehole was also tested. The Rn concentrations in groundwaters from boreholes of the study site ranged from 192 to 1597 Bq L−1. The Rn exhalation rates of selected samples of granite and micaschist were determined from laboratory experiments. The results yielded fluxes varying from 0.5 to 1.3 mBq m−2  s−1 in granite and from 0.5 to 0.9 mBq m−2  s−1 in micaschists. Pumping tests were performed in the studied boreholes to estimate the transmissivity and calculate the equivalent hydraulic aperture of the fractures. Transmissivities ranged from 10−5 to 10−3  m2  s−1. Using the cubic law, hydraulic equivalent fracture apertures were calculated to be in the range of 0.5–2.3 mm.To gain a better insight into the spatial variability of Rn contents in groundwater, theoretical Rn concentrations were calculated from an available simple crack model using results from field and laboratory experiments. This model gave satisfactory results for boreholes characterized by low-flow conditions, in which case, the calculated Rn contents were in the range of Rn concentrations set by the analytical uncertainty of concentrations measured in water. However, for boreholes characterized by high-flow conditions, the model underestimated the Rn concentration in groundwater. The higher the flow in the fracture, the larger the difference between calculated and measured Rn concentrations in water. These observations led to performing pumping tests to obtain a better understanding of the hydrogeological control of Rn content in water. The results clearly show an increase of Rn content in groundwater after the pumping test, which could be explained by the input of Rn-rich waters from the host matrix.

Factors controlling regional spatial distribution of 53 elements in coastal sea sediments in northern Japan: Comparison of geochemical data derived from stream and marine sediments by Atsuyuki Ohta; Noboru Imai; Shigeru Terashima; Yoshiko Tachibana; Ken Ikehara; Hajime Katayama; Atsushi Noda (357-376).
In all, 53 elements were analyzed in 1406 coastal sea sediment samples collected from an area off Hokkaido and the Tohoku region of Japan during a nationwide marine geochemical mapping project. The spatial distribution patterns of the elemental concentrations in coastal seas along with the existing geochemical maps in terrestrial areas were used to define natural geochemical background variation and mass transport processes. The terrestrial area is covered by mafic volcanic rocks and accretionary complexes associated with ophiolite, which has small amounts of felsic volcanic rocks and granite. The spatial distribution patterns of elements enriched in mafic lithologies such as Fe (Total Fe2O3) and Sc in marine environments are influenced by adjoining terrestrial materials. The spatial distribution patterns of Cr and Ni concentrations, which are highly abundant in ultramafic rocks on land, are used to evaluate the mass transport from land to the sea and the dispersive processes caused by oceanic currents. The scale of mass transport by oceanic currents occurs up to a distance of 100–200 km from the coast along the coastal areas. The regional differences of elements rich in felsic lithologies such as K (K2O), Nb and La in marine sediments are determined mainly by the relative proportion of minerals and lithic fragments enriching felsic materials to those associated with mafic materials. The spatial distribution of elemental concentration is not always continuous between the land areas and coastal sea areas. That difference is interpreted as resulting from (1) transportation of marine sediments by oceanic currents and storm waves, (2) contribution of volcanic materials such as tephra, (3) occurrence of shell fragments and foraminifera tests and (4) distribution of relict sediments of the last glacial age and early transgression age. Contamination with Cu, Zn, Cd, As, Mo, Sn, Sb, Hg, Pb and Bi was not observed in marine environments because the study area has little anthropogenic activity. Terrestrial materials are the dominant source for these metals. The Mo, Cd, Sn, Sb, Hg, Pb and Bi are abundant in silty and clayey sediments locally because of early diagenetic processes, authigenic precipitation and organic substances associated with these elements. The spatial distribution of As concentration shows exceptions: it is concentrated in some coarse and fine sands on the shelf. The enrichment is explained by adsorption of As, sourced from a coal field, to Fe hydroxide.

Chromium oxidation by manganese (hydr)oxides in a California aquifer by Kuria Ndung’u; Stephan Friedrich; Ana R. Gonzalez; A. Russell Flegal (377-381).
There are increasing concerns with elevated levels of Cr(VI) in the environment because it is a strong oxidant, corrosive, and carcinogenic. The concerns extend to the presence of Cr(VI) in many aquifers in California and elsewhere, where relatively high levels have been attributed to both industrial pollution and natural processes. The authors have, therefore, determined if natural redox processes contribute to the presence of high Cr(VI) concentrations (6–36 μg L−1) in an aquifer in central California relative to non-detectable concentrations (<0.1 μg L−1) in an adjacent aquifer. Specifically, the distribution and the redox speciation of dissolved (<0.45 μm) Cr have been compared with those of particulate Mn and Fe oxy-hydroxides in sediments, using X-ray absorption spectroscopy at the Mn and Fe L-edges. The analyses show a correlation between the presence of dissolved Cr(VI) and Mn (hydr)oxide minerals, which are the only common, naturally occurring minerals known to oxidize Cr(III) in laboratory experiments. This covariance substantiates the results of those experiments and previous field studies that indicate natural oxidation mechanisms might account for the relatively high levels of Cr(VI) in the study site, as well as for elevated concentrations in other aquifers with similar biogeochemical conditions.

Two investigated long-time stored oils, which were produced in the 1960s, show strong compositional changes compared to fresher oil samples from the same well and production zones. Asphaltenes isolated from stored and fresher-produced oil pairs show highly similar results from open-system pyrolysis. However, asphaltenes from long-time stored oils show higher reactivity compared to those from fresher oils. The study shows that differences exist in kinetic models based on asphaltenes from fresh-produced oil and those based on oil that has undergone long-term storage, and that these differences may impact geological predictions using such models. Factors controlling the chemical differences between stored and fresher asphaltenes are unclear and hard to determine, because of a broad range of factors controlling compositional differences between these oil pairs. The difference in chemical kinetics might be related to chemical storage effects, but different well-site sampling techniques between decades, or even incomplete homogenization of the long-time stored oils in barrels before sampling may also have an impact.

Characterization of manganese oxide precipitates from Appalachian coal mine drainage treatment systems by Hui Tan; Gengxin Zhang; Peter J. Heaney; Samuel M. Webb; William D. Burgos (389-399).
The removal of Mn(II) from coal mine drainage (CMD) by chemical addition/active treatment can significantly increase treatment costs. Passive treatment for Mn removal involves promotion of biological oxidative precipitation of manganese oxides (MnO x ). Manganese(II) removal was studied in three passive treatment systems in western Pennsylvania that differed based on their influent Mn(II) concentrations (20–150 mg/L), system construction (±inoculation with patented Mn(II)-oxidizing bacteria), and bed materials (limestone vs. sandstone). Manganese(II) removal occurred at pH values as low as 5.0 and temperatures as low as 2 °C, but was enhanced at circumneutral pH and warmer temperatures. Trace metals such as Zn, Ni and Co were removed effectively, in most cases preferentially, into the MnO x precipitates. Based on synchrotron radiation X-ray diffraction and Mn K-edge extended X-ray absorption fine structure spectroscopy, the predominant Mn oxides at all sites were poorly crystalline hexagonal birnessite, triclinic birnessite and todorokite. The surface morphology of the MnO x precipitates from all sites was coarse and “sponge-like” composed of nm-sized lathes and thin sheets. Based on scanning electron microscopy (SEM), MnO x precipitates were found in close proximity to both prokaryotic and eukaryotic organisms. The greatest removal efficiency of Mn(II) occurred at the one site with a higher pH in the bed and a higher influent total organic C (TOC) concentration (provided by an upstream wetland). Biological oxidation of Mn(II) driven by heterotrophic activity was most likely the predominant Mn removal mechanism in these systems. Influent water chemistry and Mn(II) oxidation kinetics affected the relative distribution of MnO x mineral assemblages in CMD treatment systems.

Discriminating between background and mine-impacted groundwater at the Phoenix mine, Nevada USA by Andy Davis; Kirk Heatwole; B. Greer; R. Ditmars; R. Clarke (400-417).
Differentiating between mineralized and non-mineralized background groundwater chemistry at a mine site can be challenging if there is an overprint of past and/or current mining on naturally mineralized conditions. At the Phoenix mine in the Copper Canyon mining district of Nevada, quantile–quantile H+/SO4 plots were used to segregate four wells clearly impacted by historical mining activity. The mineralogy of rock at the elevation of the well screen interval was used to partition the 53 remaining wells into mineralized and non-mineralized populations. For each class, groundwater chemistry was examined to identify if SO4 and H+ trends were stable (unimpacted) or increasing (impacted). Then each well was mapped as one of four resulting groundwater types across the mine site, defining the spatial extent of the different groups. Several groundwater regions were identified. A group of mineralized, mine-impacted wells (Type II) are located in Philadelphia Canyon adjacent to the Cu leach facility, with the anthropogenically impacted area bounded by several hydrologically downgradient, mineralized, unimpacted wells (Type I) to the south and east. There is a set of non-mineralized, impacted wells (Type III) downgradient from the tailings pond facility, where a historical release of SO4 is apparent in the well record. However, in some downgradient wells the tailings pond pump-back mitigation system has resulted in recovery of the groundwater quality to a non-mineralized background condition. Finally, in the vicinity of the Reona heap leach pad, there is a group of non-mineralized, unimpacted wells (Type IV). Not surprisingly, most mineralized wells (Types I and II) are located in or near mined areas, while non-mineralized wells (Types III and IV) tend to be in the southern portion of the facility in the alluvia of Buffalo and Reese River valleys.Once the well population was categorized, Phx-F was identified with extremely mineralized background groundwater (pH = 3.8; SO4  = 4600 mg/L; As = 0.056 mg/L; Cu = 72 mg/L; Zn = 61 mg/L), which required further analysis. The elevation of the Phx-F screen interval had up to 5% jarosite, with mineralized fracture zones mapping directly to cavities in the adjacent Iron Canyon mine that originally contained a saturated ferrous sulfate solution, explaining the anomalous chemistry in this well.This study demonstrates that an empirical approach using mining history, deposit geology, hydrogeology, well screen elevation mineralogy, groundwater chemistry trends, statistical analysis, and geochemical modeling provides a technique to distinguish among different classes of groundwater at the operating Phoenix mine site. Such an approach (or a derivative) may prove useful at other mines with different mineralogies and hydrogeologic settings.

Experimental study of cadmium interaction with periphytic biofilms by O.S. Pokrovsky; A. Feurtet-Mazel; R.E. Martinez; S. Morin; M. Baudrimont; T. Duong; M. Coste (418-427).
This study addresses the interaction of Cd with natural biofilms of periphytic diatoms grown during different seasons in metal-contaminated and metal-non-contaminated streams, along a tributary of the Lot River, France. Specifically, it aims to test whether the biofilms from contaminated sites have developed a protective mechanism due to high Cd exposure. Towards this goal, reversible adsorption experiments on untreated biofilms were performed in 0.01 M NaNO3 with a pH ranging from 2 to 8, Cd concentration from 0.5 to 10,000 μg/L and exposure time from 1 to 24 h. Two types of experiments, pH-dependent adsorption edge and constant-pH “Langmuirian”-type isotherms were conducted. Results were adequately modeled using a Linear Programming Model. It was found that the adsorption capacities of natural biofilm consortia with respect to Cd do not depend on season and are not directly linked to the growth environment. The biofilms grown in non-contaminated (4.6 ppb Cd in solid) and contaminated (570 ppb Cd in solid) settings exhibit similar adsorption capacities in the Cd concentration range in solution of 100–10,000 μg/L but quite different capacities at low Cd concentration (0.5–100 μg/L); unexpectedly, the non-contaminated biofilm adsorbs approximately 10 times more Cd than the contaminated one. It is therefore possible that the strong low-abundant ligands (for example, phosphoryl or sulfhydryls) are already metal-saturated on surfaces of biofilm grown in the contaminated site whereas these sites are still available for metal adsorption in samples grown in non-contaminated sites.

Gases in Taiwan mud volcanoes: Chemical composition, methane carbon isotopes, and gas fluxes by Hung-Chun Chao; Chen-Feng You; Chih-Hsien Sun (428-436).
Mud volcanoes are important pathways for CH4 emission from deep buried sediments; however, the importance of gas fluxes have hitherto been neglected in atmospheric source budget considerations. In this study, gas fluxes have been monitored to examine the stability of their chemical compositions and fluxes spatially, and stable C isotopic ratios of CH4 were determined, for several mud volcanoes on land in Taiwan. The major gas components are CH4 (>90%), “air” (i.e. N2  + O2  + Ar, 1–5%) and CO2 (1–5%) and these associated gas fluxes varied slightly at different mud volcanoes in southwestern Taiwan. The Hsiao-kun-shui (HKS) mud volcano emits the highest CH4 concentration (CH4  > 97%). On the other hand, the Chung-lun mud volcano (CL) shows CO2 up to 85%, and much lower CH4 content (<37%). High CH4 content (>90%) with low CO2 (<0.2%) are detected in the mud volcano gases collected in eastern Taiwan. It is suggestive that these gases are mostly of thermogenic origin based on C1 (methane)/C2 (ethane) + C3 (propane) and δ 13CCH4 results, with the exception of mud volcanoes situated along the Gu-ting-keng (GTK) anticline axis showing unique biogenic characteristics. Only small CH4 concentration variations, <2%, were detected in four on-site short term field-monitoring experiments, at Yue-shi-jie A, B, Kun-shui-ping and Lo-shan A. Preliminary estimation of CH4 emission fluxes for mud volcanoes on land in Taiwan fall in a range between 980 and 2010 tons annually. If soil diffusion were taken into account, the total amount of mud volcano CH4 could contribute up to 10% of total natural CH4 emissions in Taiwan.

A new groundwater radiocarbon correction approach accounting for palaeoclimate conditions during recharge and hydrochemical evolution: The Ledo-Paniselian Aquifer, Belgium by P.C. Blaser; M. Coetsiers; W. Aeschbach-Hertig; R. Kipfer; M. Van Camp; H.H. Loosli; K. Walraevens (437-455).
The particular objective of the present work is the development of a new radiocarbon correction approach accounting for palaeoclimate conditions at recharge and hydrochemical evolution. Relevant climate conditions at recharge are atmospheric pCO2 and infiltration temperatures, influencing C isotope concentrations in recharge waters. The new method is applied to the Ledo-Paniselian Aquifer in Belgium. This is a typical freshening aquifer where recharge takes place through the semi-confining cover of the Bartonian Clay. Besides cation exchange which is the major influencing process for the evolution of groundwater chemistry (particularly in the Bartonian Clay), also mixing with the original porewater solution (fossil seawater) occurs in the aquifer. Recharge temperatures were based on noble gas measurements. Potential infiltration water compositions, for a range of possible pCO2, temperature and calcite dissolution system conditions, were calculated by means of PHREEQC. Then the sampled groundwaters were modelled starting from these infiltration waters, using the computer code NETPATH and considering a wide range of geochemical processes. Fitting models were selected on the basis of correspondence of calculated δ 13C with measured δ 13C. The 14C modelling resulted in residence times ranging from Holocene to Pleistocene (few hundred years to over 40 ka) and yielded consistent results within the uncertainty estimation. Comparison was made with the δ 13C and Fontes and Garnier correction models, that do not take climate conditions at recharge into account. To date these are considered as the most representative process-oriented existing models, yet differences in calculated residence times of mostly several thousands of years (up to 19 ka) are revealed with the newly calculated ages being mostly (though not always) younger. Not accounting for climate conditions at recharge (pCO2 and temperature) is thus producing substantial error on deduced residence times. The derived 14C model ages are correlated with He concentrations measured in the groundwater of the aquifer. The obtained residence times show a gap between about 14 and 21 ka indicating possible permafrost conditions which inhibited any groundwater recharge.

In this study, the geochemistry and origin of natural gas and formation waters in Devonian age organic-rich shales and reservoir sandstones across the northern Appalachian Basin margin (western New York, eastern Ohio, northwestern Pennsylvania, and eastern Kentucky) were investigated. Additional samples were collected from Mississippian Berea Sandstone, Silurian Medina Sandstone and Ordovician Trenton/Black River Group oil and gas wells for comparison. Dissolved gases in shallow groundwaters in Devonian organic-rich shales along Lake Erie contain detectable CH4 (0.01–50.55 mol%) with low δ13C–CH4 values (−74.68 to −57.86‰) and no higher chain hydrocarbons, characteristics typical of microbial gas. Nevertheless, these groundwaters have only moderate alkalinity (1.14–8.72 meq/kg) and relatively low δ13C values of dissolved inorganic C (DIC) (−24.8 to −0.6‰), suggesting that microbial methanogenesis is limited. The majority of natural gases in Devonian organic-rich shales and sandstones at depth (>168 m) in the northern Appalachian Basin have a low CH4 to ethane and propane ratios (3–35 mol%; C1/C2  + C3) and high δ13C and δD values of CH4 (−53.35 to −40.24‰, and −315.0 to −174.6‰, respectively), which increase in depth, reservoir age and thermal maturity; the molecular and isotopic signature of these gases show that CH4 was generated via thermogenic processes. Despite this, the geochemistry of co-produced brines shows evidence for microbial activity. High δ13C values of DIC (>+10‰), slightly elevated alkalinity (up to 12.01 meq/kg) and low SO4 values (<1 mmole/L) in select Devonian organic-rich shale and sandstone formation water samples suggest the presence of methanogenesis, while low δ13C–DIC values (<−22‰) and relatively high SO4 concentrations (up to 12.31 mmole/L) in many brine samples point to SO4 reduction, which likely limits microbial CH4 generation in the Appalachian Basin. Together the formation water and gas results suggest that the vast majority of CH4 in the Devonian organic-rich shales and sandstones across the northern Appalachian Basin margin is thermogenic in origin. Small accumulations of microbial CH4 are present at shallow depths along Lake Erie and in western NY.

Abiotic interactions between natural dissolved organic matter (NDOM) and carbonate aquifer rock may be controlling factors of biogeochemical processes and contaminant fate in carbonate aquifer systems. The importance and effects of these interactions were examined using batch adsorption experiments of soil NDOM and representative carbonate sorbents from the Floridan Aquifer. Adsorption of NDOM carbon to aquifer rocks was well-described using a modified linear model and was mostly reversible. Significant adsorption was observed at higher NDOM concentrations, while the release of indigenous organic matter from the rocks occurred at lower concentrations. Longer interaction periods led to more adsorption, indicating that adsorption equilibrium was not achieved. For relatively pure carbonate rock samples, sorbent surface area was found to be the most important controlling factor of adsorption, whereas the presence of indigenous organic matter and subdominant mineral phases were more important, when they occurred. Preferential adsorption of a high over low molecular weight and humic over fulvic components of NDOM onto carbonate sorbents was detected using liquid size exclusion chromatography and excitation–emission fluorometry, respectively. The presence of NDOM inhibited mineral dissolution, though this inhibition was not proportional to NDOM concentration as surface area and mineralogy of carbonate sorbents played additional roles. Though the NDOM–carbonate rock adsorption mechanism could not be completely determined due to the heterogeneity and complexity of NDOM and sorbent surfaces, it is speculated that both rapid and weak outer-sphere bonding and stronger but slower hydrophobic interaction occur. These results have important implications for groundwater quality and hydrogeologic projects such as aquifer storage and recovery.

This study evaluates the effects of deforestation and land-use change, as compared to natural controls, on stream water chemistry in the Subandean Amazon. Dissolved major and trace elements were determined near the stream outlet of 48 independent watersheds with varying morphology, bed rock composition and intactness of forest cover (pristine to highly exploited). Geomorphological characteristics were derived from a digital elevation model, geological formations from digitalized maps and forest cover from digital classification of SPOT satellite images. Partial least square regression and multiple linear regression showed that watershed average elevation, which ranged between 396 and 1649 m, was the strongest control on stream water chemistry, explaining >70% of the variation in K and a considerable part also for Mn, U, Mg and HCO3 with near exponential concentration increases down the altitude gradient. Forest cover, which ranged between 7% and 99%, correlated strongly with average elevation (Spearman correlation coefficient, rs  = 0.8), but had no statistically significant impact on stream solute concentrations. Thus, in the studied Subandean region, watershed scale deforestation has not resulted in measurable impacts on stream water chemistry which is dominated by the spatial variation in natural controls.

Investigation was carried out to characterize the sulfidic mine spoils in the surface layer of a spontaneously combusting waste rock stockpile. The objective was to assess its potential impacts on acid mine drainage generation. The results show that there were substantial amounts of elemental S and various sulfate minerals in the weathered materials, indicating the occurrence of significant sulfide mineral oxidation in the investigated spontaneously combusting mine spoils (SCMS). It is likely that the surfacially-occurring elemental S was derived from the deeper waste rock layers experiencing spontaneous combustion under limited aeration conditions. The significantly higher acidity, EC and SO 4 2 - concentration in the SCMS, relative to the Non-SCMS suggest that spontaneous combustion is a much faster and more powerful process driving sulfide-derived acid generation, compared to microbially-catalyzed oxidation of sulfide minerals. The export of acid sulfate materials from the spontaneously combusting waste rock stockpile not only generates severe acid mine drainage but could also act as an inducer for biologically-catalyzed oxidation of newly exposed sulfide minerals in the areas surrounding the stockpile.