Hydrochemical models of the sulphidic tailings dumps at Matchless (Namibia) and Selebi-Phikwe (Botswana)

Abstract

The sulphidic tailings dumps at Matchless (Namibia) and Selebi-Phikwe (Botswana) are located in a similar semiarid environment but have a contrasting mineralogical composition. The Matchless tailings are pyrite-rich, whereas the Selebi-Phikwe tailings are dominated by pyrrhotite. Hydrochemical models are established with computer codes for water-balance, sulphide oxidation rate and hydrochemical equilibrium calculations. The data input is based on detailed mineralogical, chemical and kinetic investigations carried out on the core of boreholes drilled in 2000 and 2003. The oxidation of pyrrhotite proceeds at a much faster rate than the oxidation of pyrite. The PYROX code, which is used for kinetic calculations, can take these differences into account by applying different oxide-coating diffusion coefficients (D2) for pyrrhotite and pyrite. Humidity-cell testing is widely used to predict the post-mining composition of drainage water in humid climates. However, the semiarid conditions at Matchless and Selebi-Phikwe only allow a minimal water flux within the dump. Under such conditions, humidity-cell testing is likely to overestimate the seepage-water pH. This is suggested by the hydrochemical equilibrium calculations for the post-mining period at Selebi-Phikwe, which predict a seepage-water pH about one unit lower than the pH at the end of the 26-weeks humidity-cell testing period. The acidity of the seepage water can be reduced by about half a pH unit, if an oxygen barrier below the evaporation zone is installed. A clay layer 0.5 m thick covered by >1.5 m tailings represents the optimal design for a wet barrier. All three computer codes used for water-balance calculations (HELP3, UNSAT-H and HYDRUS-1D), predict >85% average water saturation for such a layer, which diminishes the diffusion of oxygen into the pile and production of SO4−2 and H+. The alternative design for a dry barrier consists of a vegetated silt layer 1 m thick on top of the tailings. This barrier does not significantly influence the diffusion of oxygen although it reduces the net infiltration to ≤11 mm/year.