Characterization of preferential and matrix flow through unsaturated waste rock piles with two tracers

The spatial and temporal characteristics of preferential and matrix flow through waste rock influences seepage chemistries and this information is necessary to predict mass loadings and mineral depletion. Matrix flow can be described by the Richards equation and exhibits longer water-rock interaction times than those in preferential flow paths, which exceed flow rates expected by the experimental observer. Flow regimes through two constructed pile experiments (CPEs) (Pile 4 and 5) at the Antamina mine (Antamina, Peru) were investigated using two tracers. The CPEs have an areal footprint of 36 m (w) x 36 m (l) (height = 10 m) and are comprised of mixed waste rock. Pile 4 is composed of marble and marble/hornfels waste rock, whereas Pile 5 contains marble/hornfels (similar to Pile 4) and intrusive waste rock. Particle size distribution curves of marble to marble/hornfels waste rock reveal these materials are both coarse grained (davg = 23 ± 10 mm) and favor preferential flow regimes. Conversely, intrusive material (found in Pile 5) is significantly finer grained (davg = 1.9 ± 2.7 mm) and favours slower, matrix flow regimes. Despite the finer material in Pile 5, flow volumes from both CPE basal lysimeters reveal similar daily outflows ({\textasciitilde}2.0 mm/d) and annual evaporation estimates ({\textasciitilde} 34%). CPE-specific flow regimes were more accurately defined using two tracer solutes; bromide and chloride. A bromide solution (3 g/L; as LiBr) was applied to CPE crowns in January 2010 as a 5-year rain event at Antamina. The timing of the bromide tracer test, at the height of the wet season, was chosen to yield parameters to define the faster or preferential flow component. The chloride tracer originates from blasting residues on waste rock surfaces. The internally-distributed chloride solute and its presence at the experiment initiation (i.e., the dry season) provide parameters that capture (or characterize) a slower or matrix flow component. Both bromide and chloride breakthrough results are used in a temporal moment analysis to define effective flow parameters. Final results indicate that both Pile 4 and 5 are affected by preferential and matrix flow. However, preferential flow comprises a larger portion of the flow regime in the coarser Pile 4 than Pile 5 (i.e., 95% versus 90 %, respectively). Additionally, Pile 4 is characterized by a larger disparity between mean preferential and matrix flow residence times (i.e., 94 days and 420 days, respectively). Pile 5 displays a more homogeneous flow regime as its mean preferential and matrix flow residence times do not differ by greater than 2 times (i.e., 151 days and 287 days, respectively). The overall faster flushing of Pile 4 suggests seepage chemistries are more likely to contain lower dissolved concentrations, higher mass loadings and quicker depletion estimates than the slower flushing Pile 5. This study reveals the advantage of using applied and/or quasi-natural tracers to understand internal flow regimes in waste rock piles. Given the predominance of chloride-derived boosting chemicals in blasting emulsions, this method may be applied to many sites and waste rock piles orders of magnitude larger than the experiment used in this study.