Linked optimization model for groundwater monitoring network design

Groundwater is a major source of water supply for irrigation, industrial use, and other public consumptions due to its quality, local accessibility, and relative cost. Groundwater resource is also equally vulnerable to contamination and depletion as the surface water. The detection of these contaminants is difficult, as they are not visible as the surface water systems. The detection and monitoring of the contaminants is very much important for the prediction of the contaminant transport process, and for designing efficient remedial measures. This study involves the development of methodologies for optimal design of groundwater contamination monitoring networks for detection the contaminant movement in groundwater systems, based on the application of simulated annealing as an optimization tool, and geostatistical kriging. This methodology is developed for single and time-varying optimal network design for different management periods. A budgetary constraint is used to limit the number of monitoring wells to be installed in a particular management period. The kriging linked simulated annealing (SA)-based optimization model essentially utilizes a numerical flow and transport simulation model (MODFLOW and MT3DMS) to simulate the physical and geochemical processes. It searches for an optimal set of permissible number of monitoring wells. The specified objective function of minimizing the contaminant mass estimation error is found to be quite suitable. The methodologies developed are applied to an illustrative study area comprising of homogeneous unconfined aquifer. The performance of the methodology is evaluated for the illustrative study area and the limited evaluation results demonstrate the potential applicability of the methodologies developed