Quantifying iron removal efficiency of a passive mine water treatment system using turbidity as a proxy for (particulate) iron

Abstract

Passive treatment systems are frequently used for the removal of iron from circum-neutral, ferruginous mine discharges. Upon oxygenation, iron is quickly oxidised and precipitated, and thus predominantly present as particulate hydrous ferric oxide. The objective of this study was to quantify whether turbidity as continuously monitored via in-situ sensors could serve as a surrogate for particulate iron throughout a passive treatment system. Accordingly, a statistical relationship between turbidity and particulate iron was established for a ferruginous, circum-neutral seepage water flowing through a passive pilot plant at an abandoned mine site with c (Fepart) = 0.092 × [Turb] + 0.031. The multi-stage pilot plant was equipped with ten fixed turbidity sensors to obtain a high-resolution dataset for exploitation of the near-monocausal turbidity-iron-relationship. The statistical proxy-relationship obtained from extensive monitoring by way of linear regression allowed approximate conversion of continuously monitored turbidity to particulate iron concentration with reasonable degree of accuracy. The conversion was used to describe iron transport and removal at the 30-min logging interval of the turbidity sensors and to establish a corresponding high-resolution mass balance for the pilot plant. This high-resolution mass balance showed that simpler mass balances based on overall averages or frequent, yet episodic semi-weekly samplings tended to considerably overestimate iron mass flow and accumulation due to the temporal variability in hydraulic loading, especially during weekends. Therefore, continuous monitoring of turbidity as a proxy for iron can provide a more accurate estimate of iron removal and hence an improved assessment of the iron mass balance, especially if flow rate and/or inflow concentration are subject to fluctuations. Since iron is the decisive compliance parameter on site, the correlation can be used to greatly improve data resolution (in this study from semi-weekly to 30-min interval) and quality as well as cost efficiency of monitoring efforts.