Influence of Phased Cover Placement on the Acid-Generating Main Waste Stockpile at the Red Dog Mine, Alaska, USA

Abstract

With the weathering of iron sulfide minerals, acid rock drainage (ARD) emanates from the 60-millon tonne Main Waste Stockpile (MWS) at the Red Dog Mine. Following completion of the stockpile, a collection trench was constructed in 2012–2013 to capture and treat a portion of the ARD, and a cover system was emplaced from 2021 to 2025 to cover 90% of the stockpile. Select wells in the collection trench are associated with the different cover phases. Analysis of the water chemistry of samples collected at the wells indicates increased pH and decreased dissolved solids with each phase of the cover along with significant changes in flow and solutes such as aluminum, iron, sulfate, and zinc. Although the cover should continue to decrease ARD volume, acidity, and solute concentrations, an evaluation of historical acid production and iron sulfide consumption in the stockpile indicates a likely majority of the iron sulfide content remains available for weathering and acid production. Continued MWS ARD monitoring is necessary to evaluate the multi-year effect of the cover because of the variability of the pre-cover ARD, identification of seasonal and multi-year precipitation influences on ARD generation, and a yet to be determined influence of the cover on the volume of infiltrating precipitation.

1. Introduction

Since 1989, Teck Resources has extracted high-grade, argentiferous sphalerite-galena at the Red Dog Mine in the De Long Mountains of the western Brooks Range in Alaska (Figure 1). The acid-generating waste rock from mining of the massive sulfide deposit at the Main Pit (mined from 1989 to 2012) and the Aqqaluk Pit (mined from 2010 onward) was consolidated at the Main Waste Stockpile (MWS) adjacent to the site’s Tailings Storage Facility (TSF) (Figure 1). The completed MWS is composed of about 60 million tonnes of waste rock, covers approximately 55 hectares, and is exposed to variable weathering in this subarctic climate (Köppen classification (Dwc) with cold, long winters and warm, short summers. The weathering of iron sulfide minerals in the MWS has resulted in the release of acid rock drainage (ARD) along the toe of the MWS and into the TSF [1]. The entrance of ARD (pH as low as 1.2 and dissolved solids concentration up to 130,000 mg/L) into the TSF has influenced the mine’s ability to discharge water to the environment. To reduce ARD generation, a cover system was tested in 2016 and emplaced from 2021 to 2025 to cover 90% of the stockpile. The goal of this study is to conduct a historical evaluation of the ARD emanating from the MWS to understand the temporal changes in the water chemistry and the influence of the separate phases of the MWS cover. Identification of the cover influence and changes to ARD will assist with Teck Resource’s continual risk assessment and planning efforts to reduce ARD and its influence on ongoing operations and closure planning.

To lessen the impact of ARD on TSF water chemistry, a collection trench about 1100 m in length and up to 8 m in depth was completed along the toe of the MWS in 2013, which consisted of excavating into the weathered/fractured shale bedrock, lining with a geotextile fabric, and backfilling with coarse rock. The collection trench, which consists of four segments (Figure 2) associated with likely primary groundwater flowpaths that follow historical surface drainages [2], collects ARD from the shallow subsurface and seeps along the toe of the MWS. Each trench segment is drained by a collection pump in a shallow well (Figure 2) for diversion of the ARD to a composite ARD Tank and subsequently to water treatment facilities. Monthly water chemistry samples have been collected at each pump location and at the ARD Tank since completion of the collection trench. Prior investigations estimated peak drainage after the spring snowmelt [3] with the trench able to collect 60% to 80% of the annual ARD emanating from the pile [4].

Following completion of the MWS in 2012, Teck Resources initiated a phased placement of a stockpile cover (Figure 2 and Figure 3) that now overlays approximately 90% of the MWS. The cover consists of a geomembrane overlain by drainage aggregate and associated piping plus a top cover material from locally sourced waste rock (non-acid generating) [5]. The goal of the cover and associated protective layer (local soil and unmineralized shale and native grass seeding) is to lessen precipitation infiltration and reduce ARD generation. Monitoring of the test cover emplaced in 2016/2017 indicated that the moisture content below the cover decreased to <5% (mass/mass) compared to the above-cover moisture content, which ranged from 2% to 15% [5]. As part of a larger study into the remaining acid-generating potential of the MWS, this current study evaluated the water chemistry of the ARD collected at the MWS wells for temporal trends indicative of the effect of the phased cover placement. Additionally, the water chemistry of the combined discharge from the MWS wells (MW #1 to MW #4) at the ARD Tank was evaluated for seasonal trends indicative of the stepped influence of the cover and temporal changes indicative of the evolution of iron sulfide weathering and ARD generation.

Red Dog Geology and Acidic Drainage

The Red Dog ore is hosted in Mississippian-Pennsylvanian black shales of the Kuna Formation of the Lisburne Group [6]. The target mineralized zones are referred to as the Main, Aqqaluk, Qanaiyaq, and Paalaa deposits (associated pit names) in the Ikalukrok unit [7] (Figure 4), which is a massive and unbedded deposit with disseminated sulfide grains and aggregates [8]. The mineralized deposit consists of a sulfide-bearing barite [BaSO₄] located above interlayered zones of silicified barite, sulfidic shale, and a massive sulfide unit overlying a silicified black shale with abundant sulfide veins [7]. Major sulfide minerals, in decreasing order of abundance, include sphalerite [(Zn,Fe)S], pyrite [FeS₂ (cubic)], marcasite [FeS₂ (orthorhombic)], and galena [PbS]. The pyrite and marcasite are common throughout the ore body [8]. The sphalerite is estimated to contain up to 20% of the total Fe in the deposit, and Fe composes 6.6% of the ore body [6].

The waste rock generated from the Main Pit and Aqqaluk Pit that compose the MWS contains substantial Fe [1] and significant FeS₂ (pyrite/marcasite) content [9]. The waste rock consists of barite, black shale, and silicified barite with quartz intrusions, but sulfide minerals as veins, clasts, laminations, and/or nodules were present throughout all mined units that produced the waste rock [7]. Although substantial carbonate units are adjacent to the mined areas, carbonate minerals are limited in the ore/waste rock units because of dissolution and replacement during hydrothermal intrusions that emplaced the massive sulfide deposit (sedimentary exhalative deposit) and associated silicification [7,8]. Pyrite is the dominant FeS₂ mineral and can be present as meter-thick units and as clasts and veins [8]. The high FeS₂ zones consist of banded to nodular pyrite and marcasite, which decrease in concentration moving away from the sphalerite-galena core [7]. Pre-mining weathering of the sulfide minerals only penetrated up to 20 m below the surface, primarily driven by fracture flow in the upper shale unit, with numerous byproducts, including substantial melanterite [FeSO₄·7H₂O] and elemental sulfur [8].

The MWS contains substantial waste rock from the Siksikpuk shale, which contains >3% FeS₂ and a net neutralization potential of −91 kg CaCO₃/t [10]. Other shale waste rock, such as the Ikalukrok shale, also contains substantial FeS₂. The large content of the Siksikpuk waste rock in the MWS (19% of waste rock generated), along with the potential of large FeS₂ clasts and veins in other units, results in substantial net acid production with weathering of the MWS [11]. The discharge of ARD from the MWS is driven by infiltrating precipitation atop the pile and a potential lower zone of groundwater above and in the underlying Okpikruak shale. The bedrock Okpikruak shale was initially frozen (e.g., permafrost) prior to construction of the MWS, but the shale has thawed because of the heat generated within the stockpile from sulfide oxidation [1]. The monitoring of groundwater at the toe of the MWS has indicated the largest ARD volumes occur from May through November [3] and that much of the ARD might be a result of the spring freshet and large rainfall events (Figure 5). Additionally, there may be sufficient infiltration and percolation and upgradient groundwater to saturate the lowest layer of MWS, which may drive ARD along the weathered bedrock [3]. The goal of the cover is to reduce water content in the MWS and reduce the volume of ARD discharging to the collection trench and TSF. The potential reduction of water content can lessen the volume and acidity of ARD by limiting the oxidation reaction of solution O₂ and the sulfide minerals, given the substantially higher effectiveness of solution O₂ initiating the oxidation reaction compared to soil air O₂ [12]. The rate-limiting steps of ARD generation typically are the O₂ adsorption and oxidation rate of the pyrite surface and the oxidation rate of Fe²⁺ [13,14], where a supply of oxygenated water to the pyrite surface is necessary to sustain pyrite oxidation (e.g., O₂ dissociation adsorption), Fe²⁺ oxidation (Fe²⁺–Fe³⁺ cycling), and acid generation [15,16,17,18].

It is expected that the MWS FeS₂ have undergone weathering since initiation of the stockpile in 1989 because the MWS has been exposed to an annual precipitation range of 300 to 800 mm (1992 to 2023 [19]). The study area has experienced an increasing trend in the magnitude of precipitation and the frequency of storm events [19,20,21]. The subarctic climate has not produced an expected frozen core in the MWS because of heat generated by exothermic oxidation reactions. This condition is expected to continue with ongoing ARD generation and an assumed increasing trend for temperature and precipitation in the region [20]. Additionally, there is the possibility of a saturated zone near the base of the MWS that may have weathered the lower zone of the MWS [3], along with the underlying shale bedrock. In 2006 and 2011, an examination of groundwater temperatures along the toe of the MWS prior to construction of the collection trench indicated 9 to 11 °C groundwater flowing 6 to 12 m beneath ground surface when it was expected that this zone would be permafrost [1,3]. The infiltrated portion of the annual precipitation is unknown, but only a minor portion of the estimated 550,000 m³ of annual precipitation at the MWS [1] likely enters the MWS because of the limited infiltration capacity from the high silt and clay component of the blasted shale units. Runoff and ponding are visible at the MWS during precipitation events and snowmelt (Figure 6). An evaporation study in 2011 from May through September (non-freezing period) indicated that evaporation can exceed precipitation by 45 mm (234 mm of precipitation and 279 mm of pan evaporation [22]). Yet, percolation will occur because of preferential flowpaths [2] and event-driven infiltration, such as large storms and the spring freshet [3]. Monitoring of infiltration/percolation at multiple stations across the mine site in 2010–2011 indicated net percolation of 63 to 92 mm, or 16% to 24%, of recorded precipitation [22].

2. Materials and Methods

The emplacement of each phase of the MWS cover (Figure 2) was viewed as a distinct temporal change that would influence ARD chemistry immediately following cover emplacement. Travel time in the MWS is considered highly variable, but the volume of ARD in the collection trench responds to large precipitation events and the spring snowmelt (e.g., preferential flowpath response). The wells along the toe of the MWS (

Table 1. Cover placement, area covered, and associated collection trench well (2023 Geosyntec/Teck Red Dog Operations Construction Phasing Main Waste Stockpile, Red Dog Mine drawing).

Cover	Cover Completed	Pre-Cover Data	Area Covered (ha)	Associated Well (Figure 2)
Phase IB	2021	January 2013 to June 2021	12.95	MW #3
Phase II	2022	January 2013 to June 2022	8.38	MW #2
Phase IV	2023	January 2013 to June 2023	9.87	MW #4
Phase III 1	2024/25	All data	14.93	MW #1 2

¹ Numbering is out of order because of continued use of the Phase III area until the cover period. ² Background well: no cover placement during the study period.

and Figure 2) have sufficient water chemistry data from prior to (2013 to 2020) and during the primary cover placement period (2021 to 2024) to evaluate changes in the ARD environmental parameters and solute concentrations (Table 1). The environmental parameters and select solutes of the ARD were analyzed as pre- and post-cover data sets from January 2013 to August 2024 and for equivalent pre- and post-cover periods (e.g., two years prior to cover installation and two years post-cover for the Phase II cover area). The water level data at the individual pump locations was insufficient (data continuity issues) for flow normalization of solute concentrations or load calculations. Therefore, water chemistry and flow data from the ARD Tank (data available from 2013 through 2023) were analyzed for concentration and load trends to understand the overall system alteration with cover placement.

Water samples from the collection trench wells were collected according to the Red Dog Operations Acid Rock Drainage (ARD)/Sump Water Hut Sampling Standard Operating Procedure (Document ID 2292, Rev. 4). Sample ports are available in each pump house for monthly collection of samples (when ARD is flowing, periods of ice each year in the wells) for environmental parameters (field parameters) and solute analysis. For collection of environmental parameter data, a YSI ProDSS multiparameter probe was calibrated for the collection of pH (standard units), conductivity (mS/cm), temperature (°C), dissolved oxygen (mg/L), and oxidation-reduction potential (ORP, mV). Separate water samples were collected and preserved for submission to ACZ Laboratories, Inc. (Steamboat Springs, CO, USA) for analysis of 0.45-μm filtered anion and cation concentrations, along with analysis of dissolved solids by gravimetry (Standard Method 2540C) and acidity as CaCO₃ (Standard Method 2310B). Chloride [Cl] and sulfate [SO₄] concentrations were determined via ion chromatography (EPA National Exposure Research Laboratory (NERL) Method 300.0), and ammonia [NH₃ as N] was determined by colorimetry (EPA Method 350.1). Cation (aluminum [Al], cadmium [Cd], calcium [Ca], copper [Cu], iron [Fe], lead [Pb], magnesium [Mg], manganese [Mn], potassium [K], selenium [Se], sodium [Na], thallium [Tl], and zinc [Zn]) concentrations were determined via inductively coupled plasma atomic emission spectroscopy (EPA-NERL 200.7) or inductively coupled plasma mass spectrometry (EPA-NERL 200.87). Duplicate samples were randomly collected during each collection period to assess laboratory analysis accuracy.

Statistical Analysis

Although a test cover was emplaced atop the MWS in 2016–2017, this small cover placement did not have a substantive influence on ARD emanating from the MWS. Therefore, the phased cover period is considered to have begun with the initiation of the larger cover placements in 2021 (Figure 2), and all pre- and post-cover period comparisons are based on the separate phases of cover placement (Table 1). To determine changes in analytes between pre- and post-cover periods (entire study period and equivalent periods before and after the cover depending on the length of the post-cover period at each well), each well’s data were used to compare periods before and after a cover through summary statistics (median and standard deviation) and a statistical evaluation of sample set differences and an increase or decrease in an analyte. The Wilcoxon rank sum test was used to determine whether pre- and post-cover data were similar or different, and the Hodges-Lehmann estimator was used to determine the direction of change (increase or decrease) and an estimate of the size of the change. The Wilcoxon rank sum test is a non-parametric test that allows for comparison of two non-normal data sets for a significant difference between the sample set ranks, which allows for a comparison of the central tendency of the two data sets without assuming a specific distribution shape. A p value is presented along with the Wilcoxon rank sum test statistic to indicate the significance or lack of significance of the relation probability. The Hodges-Lehmann estimator is a nonparametric estimator of a population’s location parameter (e.g., location shift within the population) that is insensitive to non-normal distribution characteristics. Non-parametric Lowess (locally weighted smoothing) and Mann-Kendall analyses were used to evaluate trends of the ARD data. Lowess analysis employs a local regression within a selected temporal window, and the regression is weighted so that the central point gets the highest weight and points that are farther away receive less weight. The Mann-Kendall test is used to detect monotonic trends in time series data when the data is not normally distributed, and the trend is unlikely to be linear. The test determines the difference of each data point and every prior data point to determine a statistic (τ) representative of the sum of positive and negative signs of the differences (e.g., sign count). An associated p value is estimated to evaluate the likelihood of the trend. The Mann-Kendall test was used for trend analysis because of the non-normality and high variability of the data where an indication of an overall upward or downward trend over the selected period can be used to evaluate the significance of trends visible in Lowess trendlines.

3. Results

3.1. Summary Statistics and Indications of Decreased Solute Mobilization

The summary statistics of the MWS ARD indicate noticeable changes between pre- and post-cover periods for nearly all analytes (

Table 2. Pre- and post-cover, mean and standard deviation of field parameters, total dissolved solids (TDS), iron, and sulfate for acid rock drainage sampled by the collection trench wells. Cond., conductivity; ORP, oxidation reduction potential; Temp., temperature. The pre-cover periods for each monitoring well are presented in Table 1.

Well	pH	Cond. (mS/cm)	ORP (mV)	Temp. (°C)	TDS (g/L)	Fe (g/L)	SO4 (g/L)
MW #1	2.63 ± 0.39	16.00 ± 3.72	402 ± 59	10.25 ± 2.80	54.42 ± 14.65	2.29 ± 0.91	39.16 ± 12.96
MW #2:
Pre-cover	2.58 ± 0.30	16.31 ± 3.64	403 ± 40	10.16 ± 3.74	49.94 ± 12.12	2.48 ± 0.63	37.49 ± 9.29
Post-cover	2.77 ± 0.30	13.47 ± 4.17	381 ± 30	14.21 ± 2.19	39.21 ± 12.54	1.79 ± 0.87	29.71 ± 9.27
MW #3:
Pre-cover	2.27 ± 0.39	29.59 ± 31.84	458 ± 51	17.55 ± 4.52	82.27 ± 25.78	2.95 ± 0.81	60.69 ± 20.55
Post-cover	2.29 ± 0.15	21.07 ± 5.94	470 ± 48	17.85 ± 7.40	77.13 ± 29.35	2.37 ± 0.81	51.67 ± 21.21
MW #4:
Pre-cover	2.86 ± 0.35	12.65 ± 6.52	365 ± 93	14.81 ± 7.23	29.10 ± 15.36	1.44 ± 0.81	21.87 ± 10.79
Post-cover	3.12 ± 0.61	6.27 ± 4.20	320 ± 61	9.57 ± 4.28	16.08 ± 8.40	0.68 ± 0.47	10.89 ± 5.79

). The pH of the ARD collected from MW #2, #3, and #4 (wells with available pre- and post-cover periods) indicate relatively small increases in pH (mean differences of 0.02 to 0.26), although a change in pH of 0.26 represents a potential 50% decrease in acid [H⁺]. With the decrease in acidity, conductivity values between the pre- and post-cover periods decreased by 17% to 50% (mean differences of 2.84 to 8.52 mS/cm). Some of the largest changes in pre- and post-cover environmental parameters occurred at MW #4. These relatively large changes in environmental parameters and associated decreases in solute concentrations at MW #4 are likely a result of the relatively higher pH of the pre- and post-cover ARD in this trench segment and sustained oxidizing conditions (ORP mean values > +300 mV) (Table 2). Pre- and post-cover mean pH values at MW #4 were greater than all pre- and post-cover mean pH values for MW #2 and #3. The relatively higher pH during the pre-cover period at MW #4 (mean pH = 2.86) and its shift to a mean pH of 3.12 ± 0.61 influences Fe solubility in this S-rich environment in comparison to the smaller increases in the already lower pH values at MW #2 and MW #3 where both pre- and post-cover periods had a pH < 3 (Table 2). The shift in pH at MW #4 above the pH 3 threshold and sustained oxidizing conditions allows for increased precipitation of acidic hydroxysulfate minerals, such as jarosite [KFe₃(SO₄)₂(OH)₆], alunite [(KAl₃(SO₄)₂(OH)₆)], and schwertmannite [Fe₈O₈(OH)₆(SO₄)] [23,24,25,26].

The large solute concentrations of the MWS ARD correspond to the oxidative dissolution of the FeS₂ and other sulfide minerals, which releases substantial SO₄ into solution (66% to 78% of the mean TDS concentrations) and produces acid that increases the solubility of associated metals (Table 2). Correspondingly, Zn and Fe were the next greatest contributors to the TDS concentrations, respectively, and each element’s data indicates noticeable decreases in concentration following cover placement (Table 2 and

Table 3. Pre- and post-cover, mean and standard deviation of primary metal(loid) concentrations for acid rock drainage sampled by the collection trench wells. The pre-cover periods for each monitoring well are presented in Table 1.

Well	Al (mg/L)	Cd (mg/L)	Cu (mg/L)	Pb (mg/L)	Mn (mg/L)	Se (mg/L)	Zn (g/L)
MW #1	848 ± 218	65 ± 16	9.9 ± 3.2	0.45 ± 0.80	397 ± 105	0.03 ± 0.02	10.67 ± 2.92
MW #2:
Pre-cover	786 ± 223	32 ± 11	2.8 ± 1.6	0.15 ± 0.21	409 ± 136	0.03 ± 0.10	7.96 ± 1.78
Post-cover	523 ± 258	18 ± 18	1.3 ± 2.2	0.10 ± 0.32	295 ± 79	0.02 ± 0.01	5.82 ± 2.79
MW #3:
Pre-cover	1201 ± 387	67 ± 20	7.6 ± 2.9	0.03 ± 0.10	582 ± 167	0.08 ± 0.09	14.53 ± 4.34
Post-cover	964 ± 430	73 ± 38	7.1 ± 3.1	0.01 ± 0.01	465 ± 231	0.09 ± 0.04	12.57 ± 4.25
MW #4:
Pre-cover	366 ± 236	28 ± 17	5.1 ± 4.1	0.21 ± 0.18	444 ± 202	0.02 ± 0.05	4.47 ± 2.44
Post-cover	184 ± 147	10 ± 4.8	1.2 ± 1.1	0.10 ± 0.02	231 ± 126	0.01 ± 0.01	2.09 ± 1.04

). The other metal(loid) mean concentrations also decreased with cover placement except for Cd at MW #3. Changes in mean Se concentrations between the pre- and post-cover periods were relatively small (<0.1 mg/L) but these changes were relatively large in relation to the mean concentrations (<0.1 mg/L (reporting limit range of 0.00025 to 0.025 mg/L)). Corresponding to the pre- and post-cover pH, ARD data from MW #4 indicate the lowest pre- and post-cover mean concentrations of Al, Cd, Fe, Zn, and SO₄ (Table 2 and Table 3). The relatively large change in SO₄ concentrations at MW #4 between the two periods (from a mean of 21.87 g/L to 10.89 g/L) reflects the decreased weathering and solubility control of S and associated metals with the higher pH in this trench segment.

The acidity of the ARD was sufficient during the pre- and post-cover periods to mobilize Al with mean concentrations ranging from 184 to 1201 mg/L (Table 3). The presence of substantial Al in solution (>2018 USEPA aquatic life acute toxicity criteria of 4.8 mg/L), along with Fe and Mn, is indicative of the strong acid generation and mobilization of typically low solubility elements in oxidizing conditions (ORP remained positive for all but one sample from MW #4 in May 2020). The presence of substantial dissolved (0.45-µm filtration) Al concentrations indicates the likely presence of the Al³⁺ as Al(H₂O)₆³⁺ instead of the common amphoteric gibbsite [Al(OH)₃] or its amorphous form (loss of gibbsite buffering with these pH values). For pH > 2, secondary Al and Fe mineral phases, such as alunite, jarosite, schwertmannite, and ferrihydrite [Fe(OH)₃], can precipitate and sorb solution metal(loid)s [27,28,29,30]. At pH values of >4, Al phases such as gibbsite and basaluminite [Al₄(SO₄)(OH)₁₀] can similarly precipitate and sorb metals [30,31,32,33]. The settling of the resulting precipitates or floc is influenced by pH and concentrations of the mineral phases and their coagulation potential [27]. At pH thresholds of ≥3.5 (Fe) and ≥5.2 (Al), Fe³⁺ and Al³⁺ loss by precipitation can be >95% [33]. Although the loss of Al from solution typically occurs at higher pH levels compared to Fe, the growth of well-ordered gibbsite films on muscovite have been observed in solutions with a pH range of 3 to 4 [34]. With the low pH values of the pre-cover periods (average values range from 2.27 to 2.86), the formation of Fe and Al precipitates has been limited because of the high H⁺ activity. The post-cover increases in pH reduced mean Al and Fe concentrations by 20% (MW #3) to 50% (MW #4). With completion of the cover and reduction of the infiltrating precipitation, the expectation is less acid generation and greater precipitation of ochre-colored Fe oxyhydrosulfates and white precipitates characteristic of aluminum oxyhydrosulfates [35,36]. Such precipitates have been identified in past investigations near the surface and toe of the MWS, such as test pits dug in 2017 by SRK Consulting (Figure 7). Further increases in pH because of the cover placement should continue to reduce metal(loid) concentrations through less weathering and mineral precipitation and become a primary system of controlling trace element mobilization.

3.2. Pre- and Post-Cover ARD Statistical Comparison

The Wilcoxon rank sum test and Hodges-Lehman estimator values for the pre- and post-cover periods (Table 1) indicate that most of the visible changes in the mean analyte conditions/concentrations (Table 2 and Table 3) are statistically significant for each well (

Table 4. Results of two-tail Wilcoxon rank sum test with Hodges-Lehmann estimator (HL est.) comparing pre- and post-cover pH, conductivity, temperature, total dissolved solids (TDS), iron, sulfate, and zinc values for acid rock drainage collected at the trench wells. The Hodges-Lehmann statistic is the median of the pre-cover × post-cover differences. A positive statistic indicates greater values in the pre-cover population. The pre-cover periods for each monitoring well are presented in Table 1.

Well	pH	Conductivity	Temperature	TDS	Fe	SO4	Zn
MW #2:
p value	<0.01	<0.01	<0.01	<0.01	<0.01	<0.01	<0.01
HL est.	−0.25	2760	−4.4	13,600	810	8700	2530
MW #3:
p value	<0.01	<0.01	0.02	0.12	<0.01	0.21	0.01
HL est.	−0.14	5802	2.6	10,650	770	7200	3000
MW #4:
p value	0.06	<0.01	<0.01	<0.01	<0.01	<0.01	<0.01
HL est.	−0.30	6107	5.2	11,800	710	9765	2110

). Results from MW #2 indicate the most significant analyte changes between pre- and post-cover periods (p value < 0.01 for all analytes), while results from MW #3 indicate the least consistency in statistical significance, with p values ranging from <0.01 to 0.21. This difference in statistical significance in pre- and post-cover periods for the different wells indicates a lack of correlation with the length of the post-cover period since MW #3 had the longest post-cover period data set. The Hodges-Lehmann estimator results indicate the largest analyte changes in pH, conductivity, temperature, and SO₄ at MW #4, while results from MW #2 indicate the largest changes in TDS and Fe. Results from MW #3 indicate the largest change for Zn.

The results of the comparison of equivalent pre- and post-cover period analyte populations indicate no statistical significance for MW #4 analytes (

Table 5. Results of two-tail Wilcoxon rank sum test with Hodges-Lehmann estimator (HL est.) comparing equivalent pre- and post-cover periods for pH, conductivity, temperature, total dissolved solids (TDS), iron, sulfate, and zinc values for acid rock drainage collected at the trench wells. The Hodges-Lehmann statistic is the median of the pre-cover × post-cover differences. A positive statistic indicates greater values in the pre-cover population. The pre-cover periods for each monitoring well are presented in Table 1.

Well	pH	Conductivity	Temperature	TDS	Fe	SO4	Zn
MW #2:
p value	<0.01	0.08	<0.01	<0.01	<0.01	0.02	0.02
HL est.	−0.20	1617	−3.2	10,650	650	8200	1395
MW #3:
p value	<0.01	<0.01	0.24	0.02	<0.01	0.20	0.01
HL est.	−0.20	6572	1.4	17,800	820	9500	3300
MW #4:
p value	0.76	0.76	0.98	0.72	0.9	0.96	0.72
HL est.	0.08	170	−0.1	−1085	30.5	105	−201

), which had the shortest post-cover period data set (Table 1). This lack of statistical difference at MW #4 indicates that the substantial changes in mean analyte conditions/concentrations indicated by the comparison of the full pre-cover period to the post-cover period (Table 4) may mask a greater similarity in analyte conditions/conditions just prior to the cover placement and after cover placement in this trench segment. In comparison, the additional year of monitoring for the post-cover period at MW #2 strengthens the perceived difference in equivalent pre- and post-cover periods where most analytes had results that were significantly different (p values range from <0.01 to 0.08, Table 5). Although, the mix of significant and insignificant differences in equivalent pre- and post-cover periods for MW #3 (p value range of <0.01 to 0.24) suggest local differences for the trench segments. The equivalent pre- and post-cover period analysis suggests that continued monitoring should allow for increased identification of significant changes in the post-cover ARD similar to the significant differences identified in the full data pre- and post-cover period analysis (Table 4).

3.3. Temporal Trends in Dissolved Solids and pH

An examination of TDS and pH trends in ARD at each of the wells (Figure 8) indicates seasonal and annual variations with pronounced shifts with and without cover placement. Seasonal changes are visible in ARD from each well during the pre-cover period, but ARD from each trench segment represented by the wells appears to have evolved differently. From 2013 to 2021 (pre-cover period), the TDS trend in ARD from MW #1 and MW #3 indicates an increasing trend, with opposing trends from 2022 to 2024 when TDS continued to increase in MW #1 (although with highly variable values) and TDS had a substantive but variable decline (17,800 mg/L (Table 5)) at MW #3 following placement of the cover. The TDS of ARD from MW #2 and MW #4 also declined following cover placement, but the range of values is substantial and highly variable with time. The opposing trends of ARD with and without cover placement suggest a strengthening of ARD conditions in areas without a cover (MW #1), which was counteracted by placement of the cover in other areas (MW #2–4). Correspondingly, the trend in pH values at MW #1 has declined since 2022 in contrast to the increasing trend in pH at MW #2 and steady or slightly increasing trends at MW #3 and MW #4.

The results of the statistical evaluation of potential monotonic trends (Mann-Kendall test) in pH and TDS for the pre- and post-cover periods (

Table 6. Results of the Mann-Kendall test for monotonic trends for the pre- and post-cover pH and total dissolved solids (TDS) values from the collection trench wells. A negative tau (τ) value indicates a possible decreasing trend, and positive τ indicates a possible increasing trend. Associated p values indicate the potential significance of the trend. The pre-cover periods for each monitoring well are presented in Table 1.

Well (Period)	pH τ	pH p Value	TDS τ	TDS p Value
MW #1	−0.05	0.45	0.4	<0.01
MW #2:
Pre-cover	0.11	0.10	−0.11	0.09
Post-cover	0.21	0.15	−0.61	<0.01
MW #3:
Pre-cover	−0.27	<0.01	0.35	<0.01
Post-cover	−0.11	0.47	−0.07	0.66
MW #4:
Pre-cover	0.13	0.06	−0.42	<0.01
Post-cover	0.28	0.20	−0.49	0.02

) align with the perceived changes visible in the Lowess trendlines (Figure 8). At MW #1, the pH data indicate a weak (τ = −0.05) decreasing and insignificant or unlikely (p value = 0.45) trend, while the TDS data indicate a stronger (τ = 0.4) and significant or highly likely (p value < 0.01) increasing trend. At MW #2, a slight increasing trend is present in the pH data (τ = 0.11 and 0.21 for pre- and post-cover periods, respectively), although the significance of the trend decreased between the two periods (p value = 0.10 and 0.15 for the pre- and post-cover periods, respectively). Correspondingly, TDS results indicate strengthening of the decreasing trend from the pre- to post-cover period (τ = −0.11 (p value = 0.09) and −0.61 (p value < 0.01)). At MW #3, pH results indicate a significant (p value < 0.01) decreasing (τ = −0.27) trend in the pre-cover period that weakened to an insignificant (p value = 0.47) decreasing trend (τ = −0.11) with a similar change in TDS from a significant (p value < 0.01) increasing (τ = 0.35) trend to an insignificant (p value = 0.66) decreasing (τ = −0.07) trend. At MW #4, the cover has not had a substantive influence on the pH and TDS trends, but the Mann-Kendall results indicate a potential (p value = 0.2) enhancement in what was an increasing pH trend (pre-cover τ = 0.13 and post-cover τ = 0.28), which also may be visible in the TDS results where τ decreased from −0.42 to −0.49 from the pre- to post-cover period (p value < 0.01 and 0.02, respectively) (Table 6).

3.4. ARD Tank Chemistry and Inflow

An examination of the trends in pH and TDS of the MWS ARD that was composited at the ARD Tank indicates seasonal and annual variations with pronounced shifts prior to and during the phased cover placement (Figure 9). A period of low pH and high TDS occurred prior to and during the first phase of cover placement (2021 to 2022), with a subsequent improvement in water quality indicated by the substantial increase in pH and decrease in TDS, along with less variability in the values. An examination of the flow and the SO₄ and Zn loads to the ARD Tank indicates a large flushing of these solutes from the MWS in 2013, followed by a decrease and similar annual cycles with the reappearance of large loads in 2018 and 2019 (Figure 10). The relatively large SO₄ and Zn loads of 2018 occurred under low flow conditions. The SO₄ and Zn loads were smaller in 2020, with a substantial decline in flow and loads in 2021 and 2022.

With the initiation of the cover placements in 2021, the pH and TDS of the ARD Tank started to improve as indicated by the higher pH and lower TDS, SO₄, and Zn concentrations, but such conclusions may be premature given the multi-year flux of ARD that is observable in the pre-cover period (Figure 9 and Figure 10). A Mann-Kendall trend comparison of the ARD Tank data indicates that the pre-cover period (pre-July 2021) had a pH trend of −0.03 (p value = 0.59) and a post-cover pH trend of 0.32 (p value < 0.01) with a pre-cover TDS trend of −0.11 (p value = 0.05) and a post-covered TDS trend of −0.39 (p value < 0.01). These trend statistics indicate a statistically significant improvement in pH and TDS at the ARD Tank with initiation of the cover placement, which corresponds to decreases in solute concentrations/loads, such as those seen with SO₄ and Zn, the two primary solutes in the ARD. The weathering of iron sulfide minerals can produce large amounts of insoluble and slightly soluble S and Fe-bearing secondary phases consisting of various ferric (oxyhydr)oxides (e.g., ferrihydrite [Fe(OH)₃] and goethite [α-FeO(OH)]), metal sulfates, and hydroxysulfates [38] that may be found as rims/rinds or adjacent to the sulfide mineral depending on the solution environment [39,40,41,42,43]. Such mineral precipitates are present at and near the surface at the MWS and have been described as dehydration or saturation products [37] in areas where the ARD is exposed to the atmosphere (Figure 7). The formation of such minerals likely will be enhanced with covering of the MWS because of decreased infiltration, which is still sufficient to generate ARD, but the decreased solution volume allows for greater precipitation of such minerals.

3.5. Estimated Acid Generation and Iron Sulfide Consumption

With the infiltration of oxygenated water into the MWS, the FeS₂ reacted with available oxygen to produce acid (Equation (1)). Additional acid generation can occur with the cycling of Fe²⁺/Fe³⁺ [44] if the Fe remains available. With greater progression of the oxidative dissolution of FeS₂ and accumulation of solution Fe³⁺, acid generation can be 8× the initial reaction with the continued presence of FeS₂ and available Fe³⁺ (Equation (2)). Often, the reactions are combined to present an expected overall acid generation (Equation (3)) for strongly acid-generating rock. Oxidative dissolution of FeS₂ has various rate limiting/enhancing factors, such as the formation of S intermediaries and Fe (oxyhydr)oxide layers (rims), which can result in decreased acid generation but also provide the e⁻ carrier for Fe²⁺/Fe³⁺ cycling [45,46,47]. Yet, the stoichiometry of the basic acid-generating equations (Equations (1) and (2)) provides the possible lower and upper bounds of H⁺ potential that can be used to estimate the FeS₂ consumption necessary to produce the recorded pH of the ARD captured by the collection trench. To evaluate the potential consumption of FeS₂ in the MWS, the pH of the ARD Tank was used to estimate the total acid generated during the period from 2013 through 2023. The monthly data of the ARD Tank pH was converted to H⁺ concentration (mol/L), and the concentration trend was integrated to estimate the total concentration for 2013–2023. The total concentration was multiplied by the total flow to the ARD Tank (daily value) to estimate a total abundance (mol). For comparison, the total abundance of SO₄ was also calculated using the same method. The SO₄ abundance provides an equivalent measure of the oxidation of sulfide minerals and the release, and oxidation, of S forms from non-sulfide minerals.

$$\mathrm{F}\mathrm{e}{\mathrm{S}}_2+{\displaystyle \frac{7}{2}}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}={\mathrm{F}\mathrm{e}}^{2+}+2{\mathrm{S}{\mathrm{O}}_4}^{2-}+2{\mathrm{H}}^{+}$$

(1)

$$\mathrm{F}\mathrm{e}{\mathrm{S}}_2+14{\mathrm{F}\mathrm{e}}^{3+}+8{\mathrm{H}}_2\mathrm{O}=15{\mathrm{F}\mathrm{e}}^{2+}+2{\mathrm{S}{\mathrm{O}}_4}^{2-}+16{\mathrm{H}}^{+}$$

(2)

$$\mathrm{F}\mathrm{e}{\mathrm{S}}_2+{\displaystyle \frac{15}{4}}{\mathrm{O}}_2+{\displaystyle \frac{7}{2}}{\mathrm{H}}_2\mathrm{O}=\mathrm{F}\mathrm{e}{\left(\mathrm{O}\mathrm{H}\right)}_3+2{\mathrm{S}{\mathrm{O}}_4}^{2-}+4{\mathrm{H}}^{+}(\mathrm{o}\mathrm{r} 2{\mathrm{H}}_2{\mathrm{S}\mathrm{O}}_4)$$

(3)

The MWS is estimated to contain 60 million tonnes of waste rock with an FeS₂ range of 0 to 35 wt %, a median concentration of 3.7 wt %, and an average concentration of 7.7 wt % [10]. These results compare to a total S median concentration of 8.8 wt %, an average concentration of 8.4 wt %, a sulfide S median concentration of 8.1 wt %, and an average concentration of 7.8 wt % [11]. Earlier results of MWS sampling indicated a FeS₂ range of 0.7 to 9.9 wt % and average of 4.2 wt % [48]. From these past results, the MWS is assumed to have a likely FeS₂ content of 3 to 9 wt %, which indicates a possible MWS FeS₂ content of 1.8 to 5.4 million tonnes or about 15 to 45 billion mol of FeS₂ placed in the MWS from its initiation in 1989 to its completion in 2012. From Equations (1)–(3), this amount of FeS₂ has the potential to generate 30 (Equation (1)) to 720 (Equation (2)) billion mol of H⁺ with a likely potential (Equation (3)) for 60 to 180 billion mol of H⁺. Given the recorded pH and flow volume (about 3.1 million m³) of ARD from the collection trench to the ARD Tank, it is estimated that a total of about 47.5 billion mol of H⁺ were captured from 2013 to 2023. If the collection trench is capturing 60% to 80% of the ARD emanating from the MWS [4], then the MWS has produced approximately 59 to 79 billion mol of H⁺. This H⁺ production represents about 40% to 80% of the medium H⁺ potential (Equation (3)) for an assumed 5% to 7% FeS₂ content (middle of the FeS₂ range) in the MWS (

Table 7. Estimates of iron sulfide (FeS₂) and hydrogen ion (H⁺) availability, production, and/or consumption with weathering of the Main Waste Stockpile (MWS) from 2013 through 2023 for four potential FeS₂ concentrations.

	Total ARD: 60% Captured by the Trench				Total ARD: 80% Captured by the Trench
FeS2 concentration	3%	5%	7%	9%	3%	5%	7%	9%
FeS2 (Mt) 1	1.8	3.0	4.2	5.4	1.8	3.0	4.2	5.4
FeS2 (Mmol) 1	15,000	25,000	35,000	45,000	15,000	25,000	35,000	45,000
Total flow (103 m3)	5200	5200	5200	5200	3900	3900	3900	3900
ARD H+ (Mmol) 2	79,200	79,200	79,200	79,200	59,400	59,400	59,400	59,400
Low acid production (Equation (1), 2 × H+ for each mol of FeS2)
Potential H+ (Mmol)	30,000	50,000	70,000	90,000	30,000	50,000	70,000	90,000
Remaining H+ (%)	−164	−58	−13	12	−98	−19	15	34
Medium acid production (Equation (3), 4 × H+ for each mol of FeS2)
Potential H+ (Mmol)	60,000	100,000	140,000	180,000	60,000	100,000	140,000	180,000
Remaining H+ (%)	−32	21	43	56	1	41	58	67
High acid production (Equation (2), 16 × H+ for each mol of FeS2)
Potential H+ (Mmol)	240,000	400,000	560,000	720,000	240,000	400,000	560,000	720,000
Remaining H+ (%)	67	80	86	89	75	85	89	92

¹ The available FeS₂ is an estimate of the FeS₂ emplaced in the MWS during its construction based on analysis of waste rock by SRK Consulting (2000, 2003, 2020). The available FeS₂ has not been adjusted for loss from weathering prior to 2013. ² The estimated total H⁺ in the MWS ARD from 2013 through 2023 is derived from the pH and flow recorded for the ARD Tank and the estimated 60% to 80% of the total ARD captured by the collection trench [4].

). If the MWS sulfide weathering sustained high acid production through retention of Fe at the mineral/pore water interface (Equation (2)), the H⁺ released from the MWS during this period for the middle FeS₂ range (5% to 7%) represents 11% to 20% of this larger estimate of H⁺ potential (Table 7). Given past evaluations of weathering of the MWS [9,11], the 40% to 80% production of potential H⁺ appears to be a reasonable assumption of past H⁺ production and remaining H⁺ production (20% to 60%).

There is insufficient data from the construction of the MWS and monitoring of its seepage from 1989 to 2012 to estimate the produced H⁺ during this pre-collection trench period, but consumption of the FeS₂ and production of H⁺ would have been smaller given the building of the MWS over this period (less available FeS₂) and initial infiltration/wetting of the waste rock. If the pre-collection trench period produced 50% of the H⁺ that was produced during the 2013–2023 period, the results for the likely scenario of 5% to 7% FeS₂ and 4 × H⁺ per mol of FeS₂ would adjust to a potential remaining H⁺ potential (FeS₂ content) of −19% to 36%. The −19% H⁺ potential is derived from the 5% FeS₂ with 60% trench capture (11% H⁺ potential for 80% capture), which indicates that the FeS₂ content of the MWS is >5% FeS₂ because ARD generation has continued. For comparison, an assumed 9% FeS₂ for the medium acid production (Equation (3)) indicates a remaining 67% of acid production from consideration of the 2013–2023 H⁺ estimate and 51% of the H⁺ potential after weathering of the MWS from its initiation through 2023. Therefore, the remaining acid-generating potential of MWS waste rock may still be substantial.

For comparison to the H⁺ release from the MWS, the ARD sent to the ARD Tank from 2013–2023 contained a cumulative total of about 4300 Mmol of SO₄ (0.42 Mt). Using the estimated 60% to 80% ARD captured by the collection trench, an estimated 5400 to 7200 Mmol of SO₄ were released from the MWS during this period, with an associated range of 8100 to 10,800 Mmol of SO₄ since initiation of the MWS (assuming a 1.5× multiplier, as noted above). Past investigations have estimated a total sulfide S average concentration of about 7.8 to 10.6 wt % with a total S of about 10.7 wt %, indicating that nearly all of the available S is present in sulfide minerals [10,11]. However, S content can range from negligible to >30% of waste rock depending on its formation/unit association, including high S content in waste rock generated from mining the massive sulfide bed and barite units [11]. Assuming a sulfide S concentration of 9 wt %, full S oxidation and S release as SO₄ (2× per sulfide) with the weathering of the sulfide minerals (no oxidation or solubility controls), the sulfide source could produce a total of about 168,500 Mmol of SO₄. With the relatively small amount of released SO₄ estimated for the 2013–2023 period (<5%) and the entire lifespan of 1989–2023 (<7%) in comparison to the potential SO₄ release from weathering of the sulfide minerals, the comparison of the H⁺ release and SO₄ release indicates substantial retention of SO₄ in the MWS. Retention of the SO₄ likely occurs with precipitation of sulfate minerals, such as melanterite, jarosite, schwertmannite, and/or gypsum, which can precipitate from sulfate-rich waters in the pH range of 1 to 3 and remain stable in these acidic conditions [49]. Such minerals were identified during past investigations (e.g., Figure 7) of the MWS [9,11] indicative of solubility controls on SO₄. The presence of such minerals is supportive of the assumption of sustainable greater acid generation (Equations (2) and (3)) with oxidative dissolution of the MWS FeS₂.

4. Discussion

Monitoring of ARD from the Red Dog Mine’s MWS since completion of the collection trench indicates how environmental parameters and solute concentrations have changed with the phased placement of the stockpile cover. Temporal trends of water chemistry at wells associated with each cover placement show improvements in pH and dissolved solids during post-cover periods, while data from the uncovered location shows similar or worsening conditions throughout the monitoring period (2013–2024). Statistical analysis of the ARD data indicates mostly improvements in water quality with each cover placement, but the data trends prior to cover placement and the length of the post-cover monitoring periods are influences on the statistical significance between the pre- and post-cover periods. Yet longer post-cover periods can indicate greater statistical significance in water quality improvements, which should continue to strengthen with longer periods of post-cover monitoring.

With collection and analysis of the composited ARD from the entire collection trench at the ARD Tank, the resulting pH data provide insight into historical acid generation and iron sulfide consumption within the entire MWS. Given past estimates of FeS₂ and sulfur concentrations of the waste rock, weathering of the pile during the 2013–2023 period may have produced about 40% to 80% of the potential acid generation given a possible 5% to 7% FeS₂ content and conservative (medium) estimate of acid production (Equation (3)). The inclusion of the pre-completion period of the stockpile from 1989 to 2012 for estimating acid generation from its initiation through 2023 suggests a remaining acid-generating potential of ≥36% (7% FeS₂) with the potential for up to 50% of the remaining acid-generating potential (9% FeS₂). With completion of the primary stockpile cover in 2025 (90% coverage), the volume, acidity, and solute concentrations of the ARD emanating from the MWS should continue to decline, but there is sufficient remaining FeS₂ content and potential infiltration zones to allow for continued ARD production. With the potential slowing of the ARD generation because of the cover placement and reduced infiltration, it is unlikely that the stockpile will exhaust its acid-generating capacity prior to closure of the mine site that is planned for 2031. The subsequent portions of this study will be used to further investigate the acid-generating potential of the MWS to increase our understanding of ARD release from the stockpile and inform closure planning.