Introduction

Methods for determining the hydraulic properties of mine waste materials are the same as those used in other hydrogeologic studies. Both laboratory apparatus and field techniques have a role to play in measurement of the properties of tailings and waste rock. Tailings are amenable to laboratory-scale measurements due to their grain-size characteristics, however it is important to account for the effect of the applied load as tailings undergo consolidation with on-going deposition in an active tailings storage facility. Flexible wall permeameters with the capability to vary the confining pressure are well suited for this purpose. Mill tailings deposited as a slurry have an anisotropic permeability due to depositional processes, this needs to be considered when undertaking permeameter measurements. In specialized circumstances, standpipe piezometers installed in a tailings beach have been used to estimate in-situ hydraulic conductivity using falling or rising head tests. Although not common in practice, a network of extraction wells can be used to dewater a sandy tailings deposit retained behind a tailings dam to reduce the flowability of the tailings. Efficient design of such a system requires a convention pumping test program be undertaken with one or more pumping wells and nearby monitoring wells to characterize the hydraulic response of the tailings deposit.

Approaches to reliable characterization of the hydraulic properties of waste rock piles are more challenging to implement because of the wide range in grain sizes, the interaction between matrix materials and open voids, preferential flow pathways, and the partially-saturated condition of waste rock piles. Further discussion of approaches to parameter measurement follows in subsections below.

Tailings disposal facilities

The water supply required for the first few months of ore processing is usually accumulated in the tailings storage facility prior to start-up of operations. When tailings slurry (Figure 10) is discharged into the facility, the slurry segregates with coarser-sized particles settling out in closer proximity to the discharge (spigot) point with finer-grained particles transported toward the interior of the facility. Coarser particles that settle near the spigot points eventually develop a tailings beach that stands above the water line of the supernatant pond. The finer-grained tailings deposited underwater are called slimes and consist predominantly of silt and clay-sized particles. Depending on processes that influence sediment deposition and winnowing on the beach, a variable proportion of the finer particles can be trapped above the water line and modify the hydraulic properties of the beach deposit. These deposition processes lead to the tailings forming a heterogenous deposit within the facility, which is commonly also anisotropic, with a higher horizontal than vertical hydraulic conductivity.

Figure 10 – Example of spigot discharge of tailings slurry from the embankment of a tailings storage facility.

Over time, self-weight consolidation of the tailings expels pore water from the deposit and the density of the settled tailings increases. Tailings buried at greater depths have a reduced porosity and lower hydraulic conductivity than the more recently deposited tailings. As a consequence of these processes, a general picture emerges of a tailings deposit with higher hydraulic conductivity near the spigot points, lower hydraulic conductivity in more distal locations from the spigot locations, and a decrease in hydraulic conductivity with depth. Recognition of this overall structure is important when assessing the volume of process water that might leave the facility as unrecovered seepage and enter the surrounding watershed.

At mines with a significant fraction of fine to medium-grained sand particles in the tailings, a beach deposit might have a horizontal hydraulic conductivity as high as 10^-5 m/s. In more common circumstances, the tailings grid at the mill is finer and a higher proportion of finer particles are entrained in the beach, and a typical hydraulic conductivity is 10^-8 to 10^-6 m/s. In the area of the TSF where slimes are deposited, hydraulic conductivity is usually in the range from 10^-9 to 10^-8 m/s. The numerical values reported here characterize near-surface conditions. Vertical hydraulic conductivity values are usually assumed to be 5 to 20 times smaller than horizontal hydraulic conductivity, although higher values of the anisotropic ratio can occur. At mines where a very fine grind is required to extract the target minerals in the mill, the entire tailings facility might contain slimes of very low hydraulic conductivity (10^-10 m/s). Representative seepage rates for a tailings storage facility are discussed in Section 6.

With consolidation, the porosity of a tailings deposit decreases through time as excess pore pressures are dissipated and settlement of the tailings progresses. In the literature, it is common to find the void volume of a tailings deposit expressed in terms of void ratio (e), defined as the ratio of the volume of voids to the volume of solids in a unit volume. The relationship between void ratio and porosity (θ) is θ = e / (1+e). The void ratio varies with the grain size distribution of the tailings. Therefore, a profile of void ratio versus depth will differ with location inside a tailings facility, in addition to a dependence on the time since deposition. The void ratio of tailings beach deposits is commonly observed in a range from 0.6 to 1.0 (corresponding to a porosity of 38 to 50%), while the finer tailings deposited underwater form deposits with a void ratio commonly observed in the range from 1.1 – 1.2 near surface (porosity of 52 – 55%) to around 0.8 – 0.9 at greater depths (porosity of 45 – 47%).

Waste rock stockpiles

From a hydrogeological perspective, waste rock stockpiles are more complex structures than tailings deposits. The macro-scale properties of a stockpile depend upon: (i) the relative fractions of finer material and cobble/boulder-sized rocks, and (ii) the particle size distribution within the finer size fraction. Stockpiles are built either from the bottom up as a series of thin lifts (with each lift 5 to 10 m in height) or as stockpiles of greater lift height, built by pushing the waste rock down the angle-of-repose slope that forms the perimeter of an advancing dump. In this latter case, individual lifts might be 20 – 50 m in height, and sometimes much higher than that in mountainous terrain. A key factor in both situations, but especially in the second method, is the segregation of particles that occurs during construction, with boulders accumulating toward the bottom of the stockpile and the finer sand and silt-sized particles tending to accumulate closer to the top of the lift (Figure 11). This construction method also imparts a structure to the waste stockpile, with layers forming parallel to the tip face of the stockpile that influence the spatial variation in hydraulic conductivity within the stockpile.

Figure 11 – A 15 m high waste rock test pile built by end dumping, illustrating a broad range in particle size and particle segregation on the tip face.

Over the past 15 years, a number of experimental test piles 5 to 15 m high have been instrumented to better understand physical controls on infiltration through waste rock and the geochemical controls influencing mineral weathering rates and solute release. Summary papers describing these multi-year experiments provide details on test pile construction, monitoring techniques, and interpretations of the data that were collected. These test piles have been constructed in a range of climatic settings, including the Diavik mine in northern Canada (Smith et al., 2013; Neuner et al., 2013; Bailey et al., 2016), Cluff Lake mine in northern Saskatchewan Canada (Nichol et al., 2005), Antamina mine in Peru (Vriens et al., 2019), and Grasberg mine in Indonesia (Andrina et al., 2012).

Three fundamental characteristics of waste rock stockpiles determine the manner in which water moves through these structures.

First, stockpiles have water contents less than saturation with both gravity and capillary suction exerting a control on water flow through the finer-grained materials in the stockpile. A saturated zone can form at the base of the stockpile if the original ground surface has a low hydraulic conductivity that impedes the flow moving downward into native ground. The lateral distribution and pattern of groundwater flow in any basal saturated zone reflects the influence of the original ground surface topography, with flow expected to be channeled toward and along the low points on the ground surface. This process typically controls the locations around the perimeter of the stockpile where toe seeps emerge (Figure 12). Within the stockpile, localized perched zones can also develop above low-permeability impeding layers, such as old traffic surfaces incorporated within the stockpile.

Figure 12 – Toe seepage emerging at the base of a waste rock pile.

Second, waste rock stockpiles are commonly constructed with water contents less than their residual saturation values (field capacity) upon placement in the pile. Therefore, there is a wetting up period following construction in which the initial infiltration component percolating through the pile is taken into storage before the wetting front advances farther into the pile. This process can be prolonged as each lift placed on the stockpile incorporates new material below its residual water content. It can take many years or even decades for the “first flush” of pore water to appear at the base of a large stockpile (excluding the much thinner perimeter zones of the stockpile).

Third, because of the broad range in particle size of waste rock, and particle segregation that occurs during pile construction, flow is often conceptualize as having components of both matrix flow through pathways formed by the finer grain-size fractions and preferential flow occurring through macropores and larger open void spaces between boulders. Figure 13 is a photograph of an excavation in a waste rock pile; the structure observed is characteristic of a matrix dominant system, but there are also a few zones where open void space is retained. Field monitoring programs suggest that for “soil-like piles” the majority of the infiltration passes through matrix materials with the preferential flow component activated only during periods of higher infiltration on the surface of the stockpile. This response has led researchers to develop mathematical models for infiltration through waste rock stockpiles using dual porosity type models of various degrees of sophistication; with domains representing the granular matrix, preferential flow paths, and even stagnant zones not part of the active flow regime.

Figure 13 – Excavation in a waste rock stockpile. Vertical face approximately 3 m in height. In this region of the stockpile, it is anticipated matrix flow would be dominant process, but the photo also shows in the upper center a region of open voids not infilled with finer grained materials where the saturated hydraulic conductivity is likely to be much greater.

In a relatively dry climate with low infiltration rates, the majority of the rainfall and snowmelt that infiltrates the pile moves primarily through granular matrix materials. Due to capillary effects, the matrix has a higher hydraulic conductivity at low flux rates than the coarser macropore pathways where water contents are expected to be low. As infiltration rates increase, there is proportionally increased flushing of the coarser fractions of the matrix materials. At high surface infiltration rates or when surface ponding occurs on the surface of a stockpile, water can flow more rapidly through preferential pathways that develop in the coarsest materials and perhaps, the open voids. In this circumstance, there are shorter rock-water contact times, less interaction with fine-grained materials forming the matrix pathways and the outflow at the base of the stockpile will typically have lower solute concentrations due to dilution of the matrix flow component with the more rapid flow component. This process introduces challenges in characterizing the degree of spatial variability in flushing rates within a stockpile, as the flushing of the mineral surfaces is dependent upon temporal variations in infiltration rates. A “rock-like” stockpile classified as such on the basis of the particle size distribution curve could still behave from the perspective of water flow as a “soil-like” pile in areas with a dry climate. Infiltration rates might be so low that the water is primarily restricted to pathways through the finer grained fraction that is present in the pile (Amos et al., 2015). If the coarse fraction of the waste rock dominates the structure of the stockpile with non-capillary flow governing the behavior, there can be relatively rapid percolation rates through a stockpile, on the order of meters per day, if ponding on the surface of the pile develops.

There are few comprehensive data sets available to characterize the hydrogeological properties of waste rock stockpiles. Relevant properties include: (i) the porosity of the finer grained matrix materials where capillarity influences water flow; (ii) the bulk porosity of the stockpile, which accounts for both the solids forming the matrix materials and the cobbles and boulders; (iii) the water content distribution in the matrix materials of the stockpile; (iv) the saturated hydraulic conductivity; and (v) the soil water characteristic curves of the matrix materials. Infiltration rate is also a key variable that influences solute loads at the base of the stockpile. Given the wide range in particle size distribution of waste rock at different mine sites and the multiple ways in which a stockpile is built, representative values for the site-specific hydraulic properties of stockpiles are challenging to define.

The available data indicate matrix porosity values are generally similar to values for a medium to coarse-grained sand, in the range from 0.25 to 0.35. Bulk porosity values of stockpiles are lower than this range due to the presence of large boulders that do not contribute appreciably to porosity. Field measurements suggest that in most climatic regimes, at depths below the zone of evaporative influence, matrix saturation values commonly fall in the range of 60 – 80% saturation. Water content values are anticipated to be lower where the stockpile has sufficient permeability for advective air circulation within the stockpile, as this provides a means of removing water by an evaporative transfer process. Much lower water contents are expected in “rock-like” piles where there is a smaller matrix component to retain water by capillarity.

Matrix hydraulic conductivity values at saturation often fall in the range from 10^-6 to 10^-5 m/s, reflecting the common size of matrix materials in a waste rock stockpile. Soil water characteristic curves for the matrix fraction can be measured in the lab using a Tempe cell, but it is more common in practice that the grain-size distribution curve of the material is determined and empirical equations adopted to estimate a soil water – matrix suction curve. The bulk hydraulic conductivity of a stockpile, where saturated, is several orders of magnitude greater than that of the matrix materials. Pumping tests in re-saturated waste rock backfilled in open pits indicate hydraulic conductivity estimates in the range from 10^–5 to 10^-1 m/s, with values at the higher end of the range corresponding to waste rock materials with a low fines content.

Studies of air circulation within waste rock stockpiles also provide insight to the large-scale permeability of a waste rock stockpile, either by direct estimation of air permeability from air injection, or by modeling air circulation in the stockpile and using model calibration to derive a permeability estimate. Air circulation is an important attribute when it provides oxygen re-supply to sustain sulfide mineral oxidation rates in the interior of the stockpile. Because moisture is held in the finer matrix materials, air permeability measurements generally reflect the conductivity of the open, interconnected voids, yielding higher permeability values than the finer-grained matrix material. Air permeability values for stockpiles fall within the general range of 10^-12 to 10^-8 m². For comparison, these air permeability values would indicate, under saturated conditions, bulk hydraulic conductivity values for water flow of 10^-5 to 10^-1 m/s. A number of spatially distributed air permeability tests in the waste rock test pile shown in Figure 11 yielded an average value of air permeability of 1.4 x 10^-9 m² (Amos et al., 2009). Under saturated conditions, this would correspond to a hydraulic conductivity for water of about 1 x 10^-2 m/s. This value was consistent with estimates of saturated hydraulic conductivity derived using constant head tests carried out in field permeameters two meters in height.

Tracer tests have been used to characterize velocities of water moving downward in a waste rock stockpile after initial wetting. Where the flow is restricted to matrix materials, water velocities of 1 – 10 cm/d have been inferred, although the experimental database is limited in scope. These estimates apply in areas of moderate precipitation, not arid regions. During and following high-intensity rainfall events, where preferential flow through macropore pathways can be initiated in soil-like piles, water velocities up to 1 to 2 m/d have been inferred from tracer tests. Barbour and others (2016) demonstrate the use of stable isotopes to estimate rates of water movement in a waste rock pile at a coal mine in western Canada.