Spatio-temporal assessment of groundwater depletion examining 35-year case study of Lahore (1990–2025): a machine learning approach
摘要
Groundwater resources in Punjab, Pakistan, are under increasing pressure due to over-abstraction and rapid urbanization, particularly in Lahore. The city’s groundwater table has declined at an average rate of 0.9 m per year, with a total depletion of about 40 m since 1990, and is projected to reach nearly 70 m below the 1990 level by 2030 if current trends persist. This study evaluates groundwater conditions in Lahore over a 35-year period (1990–2025) by analyzing spatio-temporal variations in Land Surface Temperature, precipitation, potential evapotranspiration, and Land Use and Land Cover, with emphasis on extreme monsoon events such as cloudbursts. Land Surface Temperature was integrated with standardized drought indices to assess groundwater storage dynamics and recharge patterns. Remote sensing datasets were obtained through Google Earth Engine, including Moderate Resolution Imaging Spectroradiometer for Land Surface Temperature anomalies, Climate Hazards Group InfraRed Precipitation with Station data for precipitation, TerraClimate for potential evapotranspiration, Gravity Recovery and Climate Experiment for groundwater storage anomalies, and Landsat for Land Use and Land Cover classification, which was validated using the Random Forest algorithm. Spatial mapping was conducted using Inverse Distance Weighting in Geographic Information System. Results indicate a 38.4% decline in groundwater levels between 1990 and 2025, largely driven by excessive extraction and reduced infiltration due to urban expansion, and increasing atmospheric demand. Although extreme monsoon events provide temporary recharge, their impact remains limited and insufficient to offset long-term depletion trends.