Conventional disposal of waste rock results in the construction of benches with interbedded fine and coarse layers dipping at the angle of repose. The waste rock benches are typically 20-meters in height and are constructed in a vertical sequence to form waste rock dumps commonly greater than 100-meters high. The interbedded structure influences the flow pathways for infiltration water within the waste rock profile. Preferential flow pathways develop when one material becomes more conductive than the surrounding material. The flow of meteoric waters through the interbedded waste rock structure is difficult to describe since the dumps are constructed above natural topography and are generally unsaturated. Two previous research studies were undertaken at the University of Saskatchewan to study end dumped waste rock piles and the relationship to preferential flow for unsaturated conditions. The first study was conducted during the excavation of a large waste rock pile at Golden Sunlight Mine in Montana (Herasymuik, 1996). Field observations showed that the waste rock pile consisted of steeply dipping fine and coarse-grained layers. The results of further laboratory analysis indicated the potential for preferential flow through the fine-grained material under conditions with negative pore-water pressures and unsaturated flow. The second study investigated the mechanism for preferential flow in vertically layered, unsaturated soil systems (Newman, 1999). The investigation included a vertical two-layer column study and a subsequent numerical modelling program showing that water prefers to flow in the finer-grained material. The preferential flow path was determined to be a function of the applied surface flux rates and the unsaturated hydraulic conductivity of the fine-grained material layer. A numerical modelling program to evaluate preferential flow was conducted for the present study in an inclined four-layer system consisting of alternating fine and coarse-grained waste rock. The numerical modelling program was undertaken using the commercial seepage software package, Seep/W, that is commonly used by geotechnical engineers. The result obtained using Seep/W showed preferential flow to occur in the fine-grained layer. However, difficulties with respect to convergence under low flow conditions with steep hydraulic conductivity functions were encountered. A comprehensive sensitivity analysis was completed to investigate the factors that influence convergence in the Seep/W model including: convergence criteria, mesh design and material properties. It was found that the hydraulic conductivity function used for the coarse-grained material was the most important factor. The problem of the steep slope for the hydraulic conductivity function specified for the coarse-grained material was solved by progressively decreasing the slope of the hydraulic conductivity function at 10-8 m/s (for applied fluxes of 10-7 m/s or less). The sensitivity analysis showed that the manipulation of the hydraulic conductivity function had insignificant changes in the flux distribution between the waste rock layers and great significance for achieving convergence. Based on the discoveries of the sensitivity analysis, a 20-meter high multi-layer waste rock profile inclined at 50º with an applied flux of 7.7e10-9 m/s equal to the annual precipitation at the Golden Sunlight Mine was successfully simulated. A parametric study was subsequently conducted for an applied flux rate of 10-5 m/s for slope heights of 1-meter to 20 meters with slope angles varying between 45º and 90º. The parametric study demonstrated that flow in a multi-layered waste rock dump is a function of inclination, contact length between the layers, and the coarse and fine-grained hydraulic properties for the waste rock. An alternative numerical modelling technique based on a modified Kisch solution was also used to investigate preferential flow. The Kisch method helped to verify and simplify the numerical problem as well as to illustrate the mechanics of preferential flow in a two-layered system. In general, commercial seepage modeling packages are powerful and useful tools that are designed to adequately accommodate a wide range of geotechnical problems. The results of this research study indicate that Seep/W may not be the best-suited tool to analyze unsaturated seepage through sloped waste rock layers. However, numerical modelling is a process and working through the process helps to enhance engineering judgment. The Seep/W model provided an adequate solution for a simplified simulation of unsaturated seepage through waste rock layers. The modified Kisch solution independently verified the solution and provided additional confidence for the results of Seep/W model.

Conventional disposal of waste rock results in the construction of benches with interbedded fine and coarse layers dipping at the angle of repose. The waste rock benches are typically 20-meters in height and are constructed in a vertical sequence to form waste rock dumps commonly greater than 100-meters high. The interbedded structure influences the flow pathways for infiltration water within the waste rock profile. Preferential flow pathways develop when one material becomes more conductive than the surrounding material. The flow of meteoric waters through the interbedded waste rock structure is difficult to describe since the dumps are constructed above natural topography and are generally unsaturated. Two previous research studies were undertaken at the University of Saskatchewan to study end dumped waste rock piles and the relationship to preferential flow for unsaturated conditions. The first study was conducted during the excavation of a large waste rock pile at Golden Sunlight Mine in Montana (Herasymuik, 1996). Field observations showed that the waste rock pile consisted of steeply dipping fine and coarse-grained layers. The results of further laboratory analysis indicated the potential for preferential flow through the fine-grained material under conditions with negative pore-water pressures and unsaturated flow. The second study investigated the mechanism for preferential flow in vertically layered, unsaturated soil systems (Newman, 1999). The investigation included a vertical two-layer column study and a subsequent numerical modelling program showing that water prefers to flow in the finer-grained material. The preferential flow path was determined to be a function of the applied surface flux rates and the unsaturated hydraulic conductivity of the fine-grained material layer. A numerical modelling program to evaluate preferential flow was conducted for the present study in an inclined four-layer system consisting of alternating fine and coarse-grained waste rock. The numerical modelling program was undertaken using the commercial seepage software package, Seep/W, that is commonly used by geotechnical engineers. The result obtained using Seep/W showed preferential flow to occur in the fine-grained layer. However, difficulties with respect to convergence under low flow conditions with steep hydraulic conductivity functions were encountered. A comprehensive sensitivity analysis was completed to investigate the factors that influence convergence in the Seep/W model including: convergence criteria, mesh design and material properties. It was found that the hydraulic conductivity function used for the coarse-grained material was the most important factor. The problem of the steep slope for the hydraulic conductivity function specified for the coarse-grained material was solved by progressively decreasing the slope of the hydraulic conductivity function at 10-8 m/s (for applied fluxes of 10-7 m/s or less). The sensitivity analysis showed that the manipulation of the hydraulic conductivity function had insignificant changes in the flux distribution between the waste rock layers and great significance for achieving convergence. Based on the discoveries of the sensitivity analysis, a 20-meter high multi-layer waste rock profile inclined at 50º with an applied flux of 7.7e10-9 m/s equal to the annual precipitation at the Golden Sunlight Mine was successfully simulated. A parametric study was subsequently conducted for an applied flux rate of 10-5 m/s for slope heights of 1-meter to 20 meters with slope angles varying between 45º and 90º. The parametric study demonstrated that flow in a multi-layered waste rock dump is a function of inclination, contact length between the layers, and the coarse and fine-grained hydraulic properties for the waste rock. An alternative numerical modelling technique based on a modified Kisch solution was also used to investigate preferential flow. The Kisch method helped to verify and simplify the numerical problem as well as to illustrate the mechanics of preferential flow in a two-layered system. In general, commercial seepage modeling packages are powerful and useful tools that are designed to adequately accommodate a wide range of geotechnical problems. The results of this research study indicate that Seep/W may not be the best-suited tool to analyze unsaturated seepage through sloped waste rock layers. However, numerical modelling is a process and working through the process helps to enhance engineering judgment. The Seep/W model provided an adequate solution for a simplified simulation of unsaturated seepage through waste rock layers. The modified Kisch solution independently verified the solution and provided additional confidence for the results of Seep/W model.

The oxidation of sulfide minerals in mine waste rock has the potential to generate acidity and contribute sulfate, metals and other trace constituents to drainage. The rate and extent to which this process occurs are dependent upon climactic conditions and the overall hydrologic, geochemical and physical properties of the waste rock. A laboratory and field-based study is currently being conducted at the Diavik Diamond Mine in the Northwest Territories, Canada, which is investigating the evolution of waste rock exposed to subaerial conditions in the continuous permafrost region. Over the course of the mine life, Diavik is expected to generate a stock pile up to 120 Mt of low-sulfide waste-rock composed primarily of granite and granite pegmatite with smaller amounts of biotite schist which occurs as xenoliths, and trace contributions from diabase dykes. Waste rock is segregated based on sulfur content into Type I (< 0.04 wt % S), Type II (0.04-0.08 wt % S) and Type III (> 0.08 wt % S) rock. The Diavik Waste Rock Research Project includes four 2 m by 2 m lysimeter experiments, two each constructed with Type I and Type III waste rock. Also constructed were two well-instrumented, 15 m high test scale waste-rock piles, one composed of Type I and one composed of Type III uncovered waste rock, and one covered test pile based on a reclamation concept which consists of a Type III waste rock core, a 1.5 m glacial till layer, and a 3 m layer of Type I waste rock. In addition, instrumentation was installed in four locations of the operational waste-rock stockpile. The geochemical differences between the Type I and Type III lysimeters and test piles is discussed to compare the non-acid generating Type I waste rock with the potentially acid-generating Type III. The effluent from the Covered test pile retained the character of the Type III waste-rock core over the course of observation producing slightly acidic drainage, possibly due to a zone of unfrozen till on the crest as a result of heat trace within the test pile. Observations from the geochemistry of the Type III waste rock will also be compared to mineralogical analysis from Type III samples collected during installation of instruments in the full scale waste-rock stockpile. Due to the low concentration of sulfide minerals, advanced techniques such as SEM and synchrotron-based analyses were employed for in-depth characterization of initial sulfide-oxidation products. SEM images and elemental mapping reveal development of reaction rims on many pyrrhotite grains, but lower instances of weathering of pyrite. Distinct zonation of weathering trends between depths within the stockpile was also absent. These observations indicate that the waste rock is in the early weathering stages may not yet be affected by the formation of permafrost. These observations act as a baseline for future studies. Correlations between the mineralogical and geochemical analyses, in addition to future monitoring and continuation of these studies, will assist in understanding the evolution of waste rock stored in a permafrost environment.

The oxidation of sulfide minerals in mine waste rock has the potential to generate acidity and contribute sulfate, metals and other trace constituents to drainage. The rate and extent to which this process occurs are dependent upon climactic conditions and the overall hydrologic, geochemical and physical properties of the waste rock. A laboratory and field-based study is currently being conducted at the Diavik Diamond Mine in the Northwest Territories, Canada, which is investigating the evolution of waste rock exposed to subaerial conditions in the continuous permafrost region. Over the course of the mine life, Diavik is expected to generate a stock pile up to 120 Mt of low-sulfide waste-rock composed primarily of granite and granite pegmatite with smaller amounts of biotite schist which occurs as xenoliths, and trace contributions from diabase dykes. Waste rock is segregated based on sulfur content into Type I (< 0.04 wt % S), Type II (0.04-0.08 wt % S) and Type III (> 0.08 wt % S) rock. The Diavik Waste Rock Research Project includes four 2 m by 2 m lysimeter experiments, two each constructed with Type I and Type III waste rock. Also constructed were two well-instrumented, 15 m high test scale waste-rock piles, one composed of Type I and one composed of Type III uncovered waste rock, and one covered test pile based on a reclamation concept which consists of a Type III waste rock core, a 1.5 m glacial till layer, and a 3 m layer of Type I waste rock. In addition, instrumentation was installed in four locations of the operational waste-rock stockpile. The geochemical differences between the Type I and Type III lysimeters and test piles is discussed to compare the non-acid generating Type I waste rock with the potentially acid-generating Type III. The effluent from the Covered test pile retained the character of the Type III waste-rock core over the course of observation producing slightly acidic drainage, possibly due to a zone of unfrozen till on the crest as a result of heat trace within the test pile. Observations from the geochemistry of the Type III waste rock will also be compared to mineralogical analysis from Type III samples collected during installation of instruments in the full scale waste-rock stockpile. Due to the low concentration of sulfide minerals, advanced techniques such as SEM and synchrotron-based analyses were employed for in-depth characterization of initial sulfide-oxidation products. SEM images and elemental mapping reveal development of reaction rims on many pyrrhotite grains, but lower instances of weathering of pyrite. Distinct zonation of weathering trends between depths within the stockpile was also absent. These observations indicate that the waste rock is in the early weathering stages may not yet be affected by the formation of permafrost. These observations act as a baseline for future studies. Correlations between the mineralogical and geochemical analyses, in addition to future monitoring and continuation of these studies, will assist in understanding the evolution of waste rock stored in a permafrost environment.

A rigorous laboratory and field study to measure and compare low sulfide waste rock and drainage characteristics at various scales has been designed and implemented. The field study was constructed at the Diavik diamond mine in the Northwest Territories, Canada. Three well-instrumented, 15 m high test piles and three sets of 2 m scale experiments were constructed from run of mine waste rock. Diavik waste rock is comprised of granite and metasedimentary biotite schist country rock. The biotite schist contains the sulfide minerals, principally pyrrhotite. Diavik segregates waste rock based on sulfur content. One test pile contains waste rock with 0.035 wt. % S, within the operational sulfur target of < 0.04 wt. % S for lower sulfur waste rock designation. The second pile contains waste rock with 0.053 wt. % S, lower than the operational sulfur target of > 0.08 wt. % S for the higher sulfur waste rock designation. The third pile contains a core of 0.082 wt. % S waste rock which is within the operational sulfur target of > 0.08 wt. % S for the higher sulfur waste rock. The third pile has been re-contoured and capped by a 1.5 m of till and 3 m of lower sulfide waste rock as per the current reclamation plan for the higher sulfide waste rock pile. The test piles were built using standard end-dumping and push-dumping methods. Instrumentation was installed at the base of each pile and on four angle of repose tip faces, as well as in the covers of the third pile. Instrumentation was selected to measure matrix flow, pore water and bulk pile geochemistry, gas-phase oxygen and carbon dioxide concentrations, temperature evolution, microbiological populations, permeability to air, and thermal conductivity, and to resolve mass and flow balances. Instruments were designed to permit measurements at multiple scales. During pile construction samples of the < 50 mm fraction of waste rock were collected. The samples were analysed for sulfur content and particle size distribution. Particle size distributions for the lower and higher sulfur waste rock are similar but the higher sulfur waste rock has a higher proportion of fines. Particle size distributions for both waste rock types suggest the piles have rock-like characteristics rather than soil-like characteristics. Sulfur concentrations vary with the scale of measurement: concentrations of smaller size fractions are higher than larger size fractions. Acid-base accounting using standard methods and site-specific mineralogical information suggest that the waste rock is acid generating. However, when acid-base accounting is compared to effluent pH and alkalinity, the data suggest these calculations may be conservative. Drainage effluent from the higher sulfide test pile was measured for field parameters (pH, Eh, alkalinity) and dissolved cations, anions and nutrients. The geochemical equilibration model MINTEQA2 was used to interpret potential geochemical controls on solution chemistry. The pH decreases to < 5, concomitant with the minimum alkalinity of < 1 mg L-1 (as total CaCO3), suggesting all available alkalinity is consumed by acid-neutralizing reactions. Sulfate concentrations reach 1995 mg L-1. Calculated saturation indices of Al (oxy)hydroxides and Al hydroxysulfate species, and pH suggest Al oxyhydroxide dissolution is buffering pH at times. Concentrations of Fe (< 0.37 mg L-1), Fe (II) and calculated saturation indices of Fe(III) (oxy)hydroxide species suggests that Fe is predominantly Fe(III) and Fe is being controlled by secondary mineral precipitation. The dissolved trace metals Mn (<19.2 mg L-1), Ni (<10.4 mg L-1), Co (<1.8 mg L-1), Zn (<0.9 mg L-1), Cd (<0.015 mg L-1) and Cu (<0.05 mg L-1) show increasing trends in the effluent water. No dissolved trace metals appear to have secondary mineral controls. Elevated SO4, Al, Fe dissolved metals Ni, Co, Zn, Cd and Cu, and depressed pH values suggest sulfide mineral oxidation is occurring in the test pile containing 0.053 wt. % S.

A rigorous laboratory and field study to measure and compare low sulfide waste rock and drainage characteristics at various scales has been designed and implemented. The field study was constructed at the Diavik diamond mine in the Northwest Territories, Canada. Three well-instrumented, 15 m high test piles and three sets of 2 m scale experiments were constructed from run of mine waste rock. Diavik waste rock is comprised of granite and metasedimentary biotite schist country rock. The biotite schist contains the sulfide minerals, principally pyrrhotite. Diavik segregates waste rock based on sulfur content. One test pile contains waste rock with 0.035 wt. % S, within the operational sulfur target of < 0.04 wt. % S for lower sulfur waste rock designation. The second pile contains waste rock with 0.053 wt. % S, lower than the operational sulfur target of > 0.08 wt. % S for the higher sulfur waste rock designation. The third pile contains a core of 0.082 wt. % S waste rock which is within the operational sulfur target of > 0.08 wt. % S for the higher sulfur waste rock. The third pile has been re-contoured and capped by a 1.5 m of till and 3 m of lower sulfide waste rock as per the current reclamation plan for the higher sulfide waste rock pile. The test piles were built using standard end-dumping and push-dumping methods. Instrumentation was installed at the base of each pile and on four angle of repose tip faces, as well as in the covers of the third pile. Instrumentation was selected to measure matrix flow, pore water and bulk pile geochemistry, gas-phase oxygen and carbon dioxide concentrations, temperature evolution, microbiological populations, permeability to air, and thermal conductivity, and to resolve mass and flow balances. Instruments were designed to permit measurements at multiple scales. During pile construction samples of the < 50 mm fraction of waste rock were collected. The samples were analysed for sulfur content and particle size distribution. Particle size distributions for the lower and higher sulfur waste rock are similar but the higher sulfur waste rock has a higher proportion of fines. Particle size distributions for both waste rock types suggest the piles have rock-like characteristics rather than soil-like characteristics. Sulfur concentrations vary with the scale of measurement: concentrations of smaller size fractions are higher than larger size fractions. Acid-base accounting using standard methods and site-specific mineralogical information suggest that the waste rock is acid generating. However, when acid-base accounting is compared to effluent pH and alkalinity, the data suggest these calculations may be conservative. Drainage effluent from the higher sulfide test pile was measured for field parameters (pH, Eh, alkalinity) and dissolved cations, anions and nutrients. The geochemical equilibration model MINTEQA2 was used to interpret potential geochemical controls on solution chemistry. The pH decreases to < 5, concomitant with the minimum alkalinity of < 1 mg L-1 (as total CaCO3), suggesting all available alkalinity is consumed by acid-neutralizing reactions. Sulfate concentrations reach 1995 mg L-1. Calculated saturation indices of Al (oxy)hydroxides and Al hydroxysulfate species, and pH suggest Al oxyhydroxide dissolution is buffering pH at times. Concentrations of Fe (< 0.37 mg L-1), Fe (II) and calculated saturation indices of Fe(III) (oxy)hydroxide species suggests that Fe is predominantly Fe(III) and Fe is being controlled by secondary mineral precipitation. The dissolved trace metals Mn (<19.2 mg L-1), Ni (<10.4 mg L-1), Co (<1.8 mg L-1), Zn (<0.9 mg L-1), Cd (<0.015 mg L-1) and Cu (<0.05 mg L-1) show increasing trends in the effluent water. No dissolved trace metals appear to have secondary mineral controls. Elevated SO4, Al, Fe dissolved metals Ni, Co, Zn, Cd and Cu, and depressed pH values suggest sulfide mineral oxidation is occurring in the test pile containing 0.053 wt. % S.

Throughout the years the mining company Boliden Mineral AB have had difficulties producing a working fine filter for their tailing pond at the Aitik copper mine. The problem with the fine filter occurs when it gets placed on the tailing pond wall. Coarse grains break down into fine grains and the entire composition of the fine filter is changed. The material that is used to produce the fine filter comes from the mine’s own waste rock supply. The primary waste rock in Aitik comprises of heterogeneous gneiss and pegmatite. To determine why the waste rock isn’t holding together well enough the mechanical strength of the rock is investigated. Huge differences for the mechanical strength both exist between the different rock types, but also in the different kinds of the gneiss. The effect of the explosions, used to mine the ore, and the crushing machine also impacts on the mechanical strength of the rock. Good mechanical strength is found in rock that has a high amount of secondary transformation like epidote and bad mechanical strength from foliated rock. To get the best mechanical properties it is suggested to exploit epidote transformed rock found in deformation zones, adjacent to the ore deposit or the pegmatite intrusions. It is also recommended to use less powerful charge when blasting rock material that will be used for production of the fine filter.

A better understanding of the hydrogeology of mine waste rock and cover systems is essential for the quantification, prediction and reduction of metals loading to the receiving environment. A series of experiments were conducted on an instrumented intermediate-scale waste rock pile at the Cluff Lake Mine in Saskatchewan to investigate the changes in flow and solute transport within coarse waste rock under three different surface conditions. Following these studies, the waste rock pile was deconstructed, structures were mapped, and samples were collected for physical characterization and pore water extraction. The internal structure of the waste rock pile was more important than the texture and topography under the free-dumped and ripped/leveled surface, while the surface condition was found to be the dominant control on spatial and temporal variability of outflow from the waste rock with the covered surface. Data from a deuterium tracer test, lysimeter outflow, and from TDR probes were used to derive estimates of the maximum and an average pore water velocity through the uncovered and the covered waste rock. An average pore water velocity through the matrix materials of the uncovered waste rock was approximately 1.5 m/yr and maximum preferential flow velocities were as high as 5 m/day. The post-cover pressure wave velocity of 0.1 to 1 m/day is inferred from outflow and TDR data, and average pore water velocities (0.39 m/y and 0.73 m/y) are calculated by the water flux and tracer methods, respectively. The distribution of the tracers in pore water and the internal structure were mapped during a detailed deconstruction of the waste rock pile and attempts were made to link the data to the spatial and temporal patterns of lysimeter outflow. The pore water chloride concentrations and the deuterium values did not provide conclusive data necessary to link the spatial and temporal patterns observed in the lysimeter hydrographs to internal structure; however, it provided insight into the internal flow mechanisms and water residence times.

A better understanding of the hydrogeology of mine waste rock and cover systems is essential for the quantification, prediction and reduction of metals loading to the receiving environment. A series of experiments were conducted on an instrumented intermediate-scale waste rock pile at the Cluff Lake Mine in Saskatchewan to investigate the changes in flow and solute transport within coarse waste rock under three different surface conditions. Following these studies, the waste rock pile was deconstructed, structures were mapped, and samples were collected for physical characterization and pore water extraction. The internal structure of the waste rock pile was more important than the texture and topography under the free-dumped and ripped/leveled surface, while the surface condition was found to be the dominant control on spatial and temporal variability of outflow from the waste rock with the covered surface. Data from a deuterium tracer test, lysimeter outflow, and from TDR probes were used to derive estimates of the maximum and an average pore water velocity through the uncovered and the covered waste rock. An average pore water velocity through the matrix materials of the uncovered waste rock was approximately 1.5 m/yr and maximum preferential flow velocities were as high as 5 m/day. The post-cover pressure wave velocity of 0.1 to 1 m/day is inferred from outflow and TDR data, and average pore water velocities (0.39 m/y and 0.73 m/y) are calculated by the water flux and tracer methods, respectively. The distribution of the tracers in pore water and the internal structure were mapped during a detailed deconstruction of the waste rock pile and attempts were made to link the data to the spatial and temporal patterns of lysimeter outflow. The pore water chloride concentrations and the deuterium values did not provide conclusive data necessary to link the spatial and temporal patterns observed in the lysimeter hydrographs to internal structure; however, it provided insight into the internal flow mechanisms and water residence times.

A better understanding of the hydrogeology of mine waste rock and cover systems is essential for the quantification, prediction and reduction of metals loading to the receiving environment. A series of experiments were conducted on an instrumented intermediate-scale waste rock pile at the Cluff Lake Mine in Saskatchewan to investigate the changes in flow and solute transport within coarse waste rock under three different surface conditions. Following these studies, the waste rock pile was deconstructed, structures were mapped, and samples were collected for physical characterization and pore water extraction. The internal structure of the waste rock pile was more important than the texture and topography under the free-dumped and ripped/leveled surface, while the surface condition was found to be the dominant control on spatial and temporal variability of outflow from the waste rock with the covered surface. Data from a deuterium tracer test, lysimeter outflow, and from TDR probes were used to derive estimates of the maximum and an average pore water velocity through the uncovered and the covered waste rock. An average pore water velocity through the matrix materials of the uncovered waste rock was approximately 1.5 m/yr and maximum preferential flow velocities were as high as 5 m/day. The post-cover pressure wave velocity of 0.1 to 1 m/day is inferred from outflow and TDR data, and average pore water velocities (0.39 m/y and 0.73 m/y) are calculated by the water flux and tracer methods, respectively. The distribution of the tracers in pore water and the internal structure were mapped during a detailed deconstruction of the waste rock pile and attempts were made to link the data to the spatial and temporal patterns of lysimeter outflow. The pore water chloride concentrations and the deuterium values did not provide conclusive data necessary to link the spatial and temporal patterns observed in the lysimeter hydrographs to internal structure; however, it provided insight into the internal flow mechanisms and water residence times. / Science, Faculty of / Earth, Ocean and Atmospheric Sciences, Department of / Graduate