The benefits of numerical groundwater modelling in resource management and scenario-testing are well known; it provides quantitative predictions of aquifer responses to stresses not yet experienced, albeit with uncertainties. Modelling is hence a widely used tool in Environmental Impact Assessment (EIA), in which prior to project commencing, the likely impacts must be assessed quantitatively to determine their significance. Based on these results mitigation measures can be proposed such that the residual impact is deemed acceptable.
At the stage of an EIA there is often very little data on which to base a model. Generally one is required to predict timescales in the order of hundreds of years with only very short-term time series data, and required to predict the response to stresses far beyond those used in the calibration. The very nature of the problems posed at EIA stage therefore render the accuracy of most modelling conducted at EIA phase severely limited. Recognising this, an appropriate model for the problems at hand can still be constructed and provide useful results.
The model results need to be seen as the first phase in an adaptive management cycle, rather than a standalone prediction which a mine can use for future operation. To strengthen the resulting predictions, the cycle in which monitoring results are used to update the model, and thus update predictions and update future requirements for monitoring repeating the cycle, needs to be entrenched into the mine phases by ensuring the recommendation as detailed in the Environmental Management Plan. Thus, what started as a useful demonstrative tool, but with large uncertainties, becomes an accurate quantitative prediction tool for operation, closure and post-closure planning.
This paper outlines a case study of a proposed open-pit zinc mine on an inselberg in South Africa, within which these themes are explored. Limited initial data was sufficient to build a useful yet simplified model. The purpose and known limitations of the model approach dictated the spatial discretisation of the model, its dimensions, and the geometry of the aquifer units, yet the simplification of the aquifer systems into the numerical model was only feasible once the complexity of the aquifer systems had been recognised, else over- or unjustified simplification is a risk.
The paper concludes with a framework for integrating the adaptive groundwater management into the mine life cycle through applying appropriate models at each phase, which would strengthen the use of groundwater models in mining.