Data-driven and numerical simulation coupling to quantify the impact of ecological water replenishment on surface water-groundwater interactions

Abstract

Ecological water replenishment (EWR) integrates surface and groundwater regulation to promote riverine baseflows and support groundwater recovery, affecting their interactions. This study introduces SWAT-LSTM-MODFLOW, an advanced SWAT-MODFLOW model incorporating LSTM networks to improve predictive accuracy in data-scarce watersheds. Applied to the Beijing section of the Yongding River Basin, the model evaluates the impact of EWR via reservoir and reclaimed water releases on groundwater recovery and SW-GW interactions. Results show that EWR enhanced groundwater levels in the short term, particularly at the mountain-plain boundary, with increases up to 5 m during high-volume replenishments. Repeated replenishments from 2019 to 2022 shifted dynamics from river seepage to increased groundwater recharge, particularly near replenishment zones. These findings highlight EWR’s role in transforming SW-GW dynamics and enhancing hydrological connectivity. This study provides a quantitative framework for assessing artificial recharge effects on groundwater and SW-GW interactions, offering a scalable methodology for similar hydrogeological conditions.