Entropy-based groundwater quality monitoring network design using a simulation-optimization approach by coupling genetic algorithm, MODFLOW and MT3DMS

Groundwater resources meet a substantial portion of Iran's water demands, making awareness of groundwater quality essential for effective resource management. However, direct water quality monitoring is both time-consuming and costly. This study aims to optimize the monitoring network for groundwater quality at the Eshtehard plain, Iran, to reduce operational costs while maximizing information gained. A simulation-optimization approach is developed based on entropy theory, using SWAT to simulate runoff and nitrate recharge to the aquifer. Groundwater flow and quality are further modeled with MODFLOW-MT3DMS, while a genetic algorithm (GA) in MATLAB determines the optimal configuration of monitoring wells. Runoff calibration in SWAT achieved a coefficient of determination of 0.85 and a Nash-Sutcliffe efficiency of 0.81. For the groundwater model, NS values of 0.99 and 0.75 were achieved for steady-state and transient calibrations, respectively. Results indicate that 7 wells, out of 19 active monitoring wells, can provide sufficient qualitative information, allowing reduced sampling effort. To validate these findings, the fuzzy C-means clustering method was applied for comparative analysis. A joint entropy-based comparison revealed that the FCM-derived network provided less information significance than the GA-optimized network, underscoring the robustness of the proposed approach for designing efficient groundwater quality monitoring networks.