An enhanced ground water level monitoring approach for sustainable water resource management

Groundwater level prediction plays an important role in rational water resource management and sustainability as traditional models often fail to capture its complex spatial and temporal variability, reducing accuracy. To address this, a new Kolmogorov-Arnold Battlefield Network (KABNet) is proposed to support effective groundwater level monitoring. The approach utilises a multi-decadal high-resolution dataset, preprocessing of missing value handling and scaling normalisation for balanced scaling and removal of magnitude bias. Feature filtering is done using the Battlefield Optimisation to retain only the most influential parameters, reducing the computational complexity. The KABNet model is optimised and trained to obtain good spatial-temporal relationships for effective groundwater prediction. Python-designed, the method outperforms traditional models on the basis of mean absolute error (MAE), root mean squared error (RMSE), and coefficient of determination (R²). The model is a good decision-support system for water resources planning and management sustainably.