The extraction of minerals through mining activities often results in the generation of acid mine drainage (AMD) due to the exposure of sulfidic minerals to oxygen and water at a rate faster than would have occurred naturally. In order to operate at best practice, the effects of mining and AMD on the natural environment need to be mitigated or prevented. In this thesis water and sediment chemistry of three historic mine sites at various stages in the post- closure process in New Zealand were assessed. Changes in water and sediment chemistry due to the precipitation of Fe, Mn and Al precipitates and the adsorption of trace elements to freshly formed precipitates were modelled in PHREEQC. The modelled results were compared to measured results in order to evaluate the effectiveness of PHREEQC in modelling post-remediation water and sediment chemistry at mining sites.

The remedial works carried out at Tui Mine have resulted in an increased pH and decreased concentration of dissolved trace elements (Fe, Cu, Pb, Zn, Cd, As) in Tui Stream and in Tunakohoia Stream. The remedial works have been more effective in mitigating the effects of AMD in Tui Stream than in Tunakohoia Stream due to the difference in the pre-remediation pH of these two streams. PHREEQC was used to show that the concentrations of Zn²⁺ and Cd²⁺ remain above ecologically relevant guidelines even after remedial works have been completed. PHREEQC was also used to show that the addition of limestone to these streams is unlikely to reduce the concentration of these two elements by either formation of mineral precipitates or adsorption to Fe (oxy)hydroxide precipitates to a concentration suitable for supporting healthy ecosystem function in the streams draining Tui Mine. The use of Tui Mine to assess the reliability of PHREEQC modelling found that it is possible to use PHREEQC to reliably model a system where the elements of concern are cationic metals such as Cu, Pb, Zn and Cd. In fact, it is possible that the remediation works carried out between 2010 and 2013 could have achieved a better outcome had PHREEQC been used to model the post-remediation water chemistry expected in the Tui and Tunakohoia Streams following the addition of limestone to the tailings dam and mine adits at Tui Mine. However, adsorption of Pb to HFO was underestimated and adsorption of As to HFO was overestimated by PHREEQC.

Historic mining operations to produce mercury, the surfacing of roads with aggregate containing Hg and As and the presence of sinter outcrops containing Hg and As have resulted in elevated concentrations of sediment-bound Hg and As, above ISQG-high guidelines, in the streams draining the Puhipuhi area. This has led to depauperate macroinvertebrate communities at the sites higher up in this catchment where sediment-bound Fe and Hg is greatest. Additionally, this has also lead to the bioaccumulation of Hg in eels which is of concern to human health, should these be consumed as part of a regular diet and the cultural value of eel harvesting for consumption. PHREEQC could be used to reliably model the speciation of cationic trace elements (Cu, Zn and Ni), the formation of Fe (oxy)hydroxides and the consequent adsorption of Cu, Zn and Ni to Fe (oxy)hydroxides. However, these elements were not the elements of concern in this catchment. PHREEQC could not be used to reliably model the speciation of dissolved Hg and the formation of methylmercury, the form of Hg which bioaccumulates, because the
process of methylation is outside the capacity of the PHREEQC model. It could not be used to reliably predict the adsorption of Hg and As to Fe (oxy)hydroxides and it could not be used to model the stability of HgS found within the quartz sinter matrix as this mineral consists of components not able to be measured in natural waters and therefore cannot be input into PHREEQC. Arsenic was associated with residual FeS₂ for which the stability again could not be reliably predicted in PHREEQC due to S²⁻ not being present in natural waters and therefore it was not able to be input into PHREEQC. Therefore, it is not recommended that PHREEQC be used to model systems where Hg and As are the elements of most environmental concern as the thermodynamic data required by PHREEQC to make accurate predictions of the behaviour of these elements is not yet known and the adsorption of As to HFO is not reliably predicted by PHREEQC.

The water chemistry of Cannel Creek prior to and following the installation of a sulfate-reducing bioreactor treatment system at Bellvue Mine was measured. The treatment system resulted in >80% less Fe, Al and Ni and a 66% less of Zn on average in Cannel Creek downstream of the AMD discharge point. Additionally, the treatment system increased the pH of AMD entering Cannel Creek sufficiently that Fe- and Al-rich precipitates were able to form and adsorption of trace elements to HFO became favourable. PHREEQC was used to develop a model predicting water chemistry in Cannel Creek following the installation of the treatment system and was validated by comparing the predicted water quality to measured water quality. The model that was developed predicted water quality in the upper reaches of the stream very well but the measured concentrations of dissolved Fe and Al deviated from those predicted in the lower reaches of the stream as there had not been sufficient time for the Fe (oxy)hydroxysulfates and Al oxides to precipitate out of solution in between sampling sites.

PHREEQC is a geochemical modelling program suitable for predicting the speciation of trace elements in AMD solutions and the attenuation of trace elements by mineral formation and adsorption to HFO. However, there are certain situations where PHREEQC cannot reliably predict trace element attenuation quantifiably. These include adsorption of dissolved Cd to HFO where solution pH is in the range of greatest change for the Cu adsorption edge, adsorption of Pb and As to HFO and methylation of Hg. In these situations PHREEQC models can provide guidance as to what processes are likely to occur but should not be depended on to set remediation targets for water quality.