Groundwater Modeling with Nonlinear Uncertainty Analyses to Enhance Remediation Design Confidence

Soil and groundwater contamination are often managed by establishing on-site cleanup targets within the context of risk assessment or risk management measures. Decision-makers rely on modeling tools to provide insight; however, it is recognized that all models are subject to uncertainty. This case study compares suggested remediation requirements using a site-specific numerical model and a standardized analytical tool to evaluate risk to a downgradient wetland receptor posed by on-site chloride impacts. The base case model, calibrated to observed non-pumping and pumping conditions, predicts a peak concentration well above regulatory criteria. Remediation scenarios are iteratively evaluated to determine a remediation design that adheres to practical site constraints, while minimizing the potential for risk to the downgradient receptor. A nonlinear uncertainty analysis is applied to each remediation scenario to stochastically evaluate the risk and find a solution that meets the site-owner risk tolerance, which in this case required a risk-averse solution. This approach, which couples nonlinear uncertainty analysis with a site-specific numerical model provides an enhanced level of knowledge to foster informed decision-making (i.e., risk-of-success) and also increases stakeholder confidence in the remediation design.