Characterisation and prediction of acid rock drainage potential in waste rock: Value of integrating quantitative mineralogical and textural measurements

The value of quantitative mineralogy and texture is often not fully appreciated within minerals applications, and particularly so in the analysis of acid rock drainage (ARD). This is partly due to the absence of a clear understanding as to why quantitative textural tools such as QEMSCAN are valuable for ARD assessment. Additionally, most ARD prediction work is performed on a coarse particle scale, for which statistical representativeness of quantitative textural and mineralogical data becomes difficult to achieve. The current study provides an application of mineralogical and textural data (Fe-sulfide liberation, association, grain size distribution) to geochemical characterisation and humidity cell tests and applies these data to the estimation of the fundamental sampling error (FSE) and mineralogical error through binomial distribution and confidence intervals for liberation data. For the ARD characterisation tests, the mineralogy was key in interpreting the classification of the samples as potentially acid forming, non-acid forming or uncertain. For ARD prediction tests, the large degree of Fe-sulfide encapsulation in intermediate weathering, slow weathering and inert minerals aided in understanding the circumneutral behaviour of potentially acid forming and uncertain samples after 60 weeks of the humidity cell testing. The particle size distribution, in conjunction with texture assessment, showed that +/- 50% of the Fesulfides were unliberated, leading to the expectation that reactivity in the humidity cells would occur predominantly as a result of mineral reaction in finer fractions. The evaluation of uncertainties suggested that one size fits all tools such as Gy's safety line or the binominal distribution approximation may not be adequate in all cases, particularly for coarse material, such as that of the parent samples or humidity cell tests. In these situations, material-specific approaches for error consideration should be taken, which in this study were the calculation of the FSE for each sample, and the plotting of size-by-size confidence intervals for the Fe-sulfide liberation data.

Automated summary

The study applies mineralogical and textural data to geochemical characterization and humidity cell tests, estimating fundamental sampling error and mineralogical error through binomial distribution and confidence intervals for liberation data. The research found that mineralogy was crucial for interpreting sample classification in ARD characterization tests, and the encapsulation of Fe-sulfides in certain minerals helped understand the behavior of samples in ARD prediction tests after humidity cell testing.