Item

Abstract

Groundwater resources worldwide are increasingly under pressure due to over-extraction and inadequate management, highlighting the urgent need for real-time monitoring and accurate forecasting to ensure sustainable usage. Although numerous studies have addressed groundwater issues, there remains a lack of clarity regarding the spatial and temporal dynamics of groundwater levels and their relationship with key water quality indicators. This study aims to fill that gap by investigating the spatial and temporal variations in Groundwater Level (GWL) along with key water quality indicators such as pH, Electrical Conductivity (EC), Total Dissolved Solids (TDS), Calcium (Ca), Magnesium (Mg), Total Hardness (TH), Bicarbonates (HCO3−), and Chlorides (Cl−). Observed data were obtained from Water and Sanitation Agency (WASA) Faisalabad observation wells, covering the period from 2013 to 2023, along with 102 uniformly distributed sample points and corresponding weather data from the TerraClimate dataset. To assess spatiotemporal variation and forecast, several machine learning (ML) and deep learning (DL) models were evaluated for their accuracy and reliability for groundwater trend projections. The eXtreme Gradient Boosting (XGBoost) model exhibited superior performance in spatial analysis, achieving the highest coefficient of determination (R2) of 0.90 and a Nash–Sutcliffe Efficiency (NSE) of 0.88, while the Long Short-Term Memory (LSTM) model excelled in temporal forecasting with an R2 = 0.91, NSE = 0.90, both accompanied by the lowest Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) values among the evaluated models. The study revealed significant seasonal variations, particularly post-monsoon increases in chemical concentrations, likely from anthropogenic and agricultural sources. These findings highlight the robustness and precision of ML & DL techniques, especially XGBoost and LSTM, in capturing complex groundwater dynamics.