Search

• Optimized water recharge using infiltration galleries prevents groundwater depletion. • Surrogate-based simulation–optimization developed the recharge management model. • Simulation model output generated the input–output patterns for surrogate models. • Bayesian and ASHA optimizations automatically selected the best surrogate model. • Planned recharge aids optimized water injection and groundwater depletion control. The Mississippi River Valley Alluvial Aquifer (MRVAA) is the main irrigation source for the Lower Mississippi River Basin. Irrigation water abstraction to meet the demands for extensive agricultural practices has contributed to groundwater depletion in this area. A managed aquifer recharge (MAR) approach has been proposed in this geographic location to minimize the impact of pumping on groundwater depletion. However, it is essential to determine the optimal amount of water to be injected through a MAR technique to reduce the decline in groundwater heads. This paper utilizes a coupled simulation–optimization (S-O) approach to estimate the optimal recharge volume into the alluvial aquifer through infiltration galleries. The aquifer processes were simulated using a physically based, three-dimensional finite-difference numerical code, MODFLOW. The MODFLOW model was calibrated and validated using the recharge rates and available groundwater head data for 26 months (27 February 2020 to 27 May 2022). The calibrated and validated models were then deployed within the coupled S-O approach to develop an aquifer recharge management model to estimate optimal groundwater recharge rates to minimize groundwater decline. Computational efficiency of the aquifer recharge management model was achieved using surrogate models that accurately reproduced the groundwater heads calculated by MODFLOW. Our evaluation demonstrates that a planned transient groundwater recharge strategy, obtained as a solution of the surrogate model based coupled S-O approach, is a useful management strategy for optimized water recharge and groundwater depletion control. This study shows the promise of the surrogate model based coupled S-O approach to potentially reduce groundwater depletion in the MRVAA by utilizing optimized recharge rates at the infiltration galleries. This work has potential applications to other aquifers and geographic locations to mitigate groundwater depletion issues due to extensive agricultural practices. [ABSTRACT FROM AUTHOR]