Characterizing future groundwater level based on precipitation and river stage considering climate change

Change in groundwater level (GWL) is an important consideration for establishing adaptive strategies against climate change to ensure the sustainability of urban social infrastructure and construction safety. In this study, long-term changes in GWL in three different regions of Seoul in Korea were predicted and characterized considering precipitation and river stage data under the projected climate change scenario of RCP 4.5. A methodology and framework for characterizing future GWL tailored to regional characteristics was established. Deep Neural Network (DNN) and Engineering Groundwater Level Prediction Model (EGPM) were employed for the prediction, utilizing precipitation and river stage data, respectively. Hydrological and geographical databases for the study regions were established and utilized for the prediction of GWL. A notable increase in future GWL compared to current level was shown, particularly pronounced in areas adjacent to rivers. By 2100, GWL in the Magok region of Seoul near Han River was projected to rise by approximately 1.8 m, while the regions farther from the river experience a moderate increase of 0.18 m. The extent and tendency of future GWL change highly depended on regional and locational characteristics. To further analyze and visualize the regional distribution of predicted GWL, the 3D flow analysis was performed, which showed how GWL changes manifest across different geographical areas.