Coulon, C, White, J, Markovich, K, Askar, A, Beal, L, Miller, S, Sigda, J, Tinklenberg, A, Ardito, C & Uliana, M 2026, 'Providing guidance for remediation planning through probabilistic modelling 
and multiple conceptual models', in AB Fourie, G Boggs, J Heyes & M Tibbett (eds), Mine Closure 2026: Proceedings of the 19th International Conference on Mine Closure, Australian Centre for Geomechanics, Perth, pp. 1-7, https://doi.org/10.36487/ACG_repo/2615_88
(https://papers.acg.uwa.edu.au/p/2615_88_Uliana/)
Abstract: Effective mine and tailings remediation planning is critical to minimise long-term environmental risks. Predictive decision-support modelling can improve remediation planning and risk-based decision-making by quantifying the uncertainty of model predictions of interest and identifying reliable remediation strategies in the face of uncertainty.  and therefore,Predictive uncertainty analysis typically focuses on quantifying and representing uncertainty in model inputs such as hydraulic properties and climate forcings, with the implicit assumption that the underlying model structure and conceptualisation are known. Although non-uniqueness in the conceptual model can strongly affect model predictions, accounting for conceptual uncertainty remains a challenge.
A decision-support tool was developed to estimate the probability of exceedance of groundwater standards by constituents of concern at a compliance boundary for multiple conceptual site models (CSMs). Each CSM could exert important controls on plume migration as well as the performance of remediation measures. The tool consists of an automated, rapid, and fully-scripted workflow which performs the following actions with little or no practitioner intervention for each plausible CSM: (1) construct a detailed site-specific groundwater flow and transport model, (2) conduct high-dimensional nonlinear history matching using an iterative ensemble smoother, (3) implement a range of possible remediation system configurations comprising source removal, a pump-and-treat system, and permeable reactive barriers, and (4) carry out a multi-objectiveoptimisation analysis using the non-dominated sorting genetic algorithm NSGA-II to explore trade-offs between economic and environmental considerations as well as reliability.
The use of an automated modelling workflow enabled the evaluation of predictive uncertainty across multiple CSMs for a large number of remediation strategies, and the identification of reliable remediation strategies considering uncertainty. The approach was implemented using freely available, open-source modelling tools and is readily adaptable to other remediation planning contexts.

Keywords: groundwater, remediation, conceptual uncertainty, optimisation under uncertainty