A decision analysis approach for optimal groundwater monitoring system design under uncertainty

Abstract. Groundwater contamination is the degradation of the natural quality of groundwater as a result of human activity. Landfills are one of the most common human activities threatening the groundwater quality. The objective of the monitoring systems is to detect the contaminant plumes before reaching the regulatory compliance boundary in order to prevent the severe risk to both society and groundwater quality, and also to enable cost-effective counter measures in case of a failure. The detection monitoring problem typically has a multi-objective nature. A multi-objective decision model (called MONIDAM) which links a classic decision analysis approach with a stochastic simulation model is applied to determine the optimal groundwater monitoring system given uncertainties due to the hydrogeological conditions and contaminant source characteristics. A Monte Carlo approach is used to incorporate uncertainties. Hydraulic conductivity and the leak location are the random inputs of the simulation model. The design objectives considered in the model are: (1) maximizing the detection probability, (2) minimizing the contaminated area and, (3) minimize the total cost of the monitoring system. The results show that the monitoring systems located close to the source are optimal except for the cases with very high unit installation and sampling cost and/or very cheap unit remediation cost.