Groundwater sustainability assessment and desertification susceptibility mapping in semi arid Bangladesh using integrated remote sensing and logistic regression modeling

Groundwater depletion poses a serious threat to water security in arid regions worldwide, risking sustainable water resources and agricultural stability. This study examines groundwater dynamics and water resource sustainability in the arid Barind Tract of Northwest Bangladesh using integrated remote sensing techniques and logistic regression modeling. It employed three key indices to assess water resource vulnerability: Normalized Difference Vegetation Index, Topsoil Grain Size Index, and Aridity Index, integrating them through logistic regression to evaluate desertification susceptibility and water sustainability. The regression model boasts an ROC value of 96.22% and R2 of 0.3893, indicating good classification performance with acceptable class variance. Results show that 82.66% of the area faces significant water resource challenges, with 6.27% (103.26 km2) at very high risk, 10.80% (177.89 km2) at high risk, and 28.17% (464.05 km2) at moderate risk. The northern regions, especially Porsha, Gomastapur, and Nachole Upazillas, are the most vulnerable to water depletion. The study recommends sustainable water management strategies, including surface water use through floating pontoons and rubber dams, emphasizing the urgent need for integrated water resource management to ensure long-term water security. Additionally, the research analyzed soil-vegetation feedback using rain use efficiency and found a negative loop in highly desertification-prone areas like Porsha and Nachole, indicating the need for regulation-based cropping practices and improved water governance in zones at risk of desertification to reduce crop-water redundancy. This research offers valuable insights for water resource planning and management in arid regions, supporting sustainable water governance and locally-led adaptation strategies for water-stressed environments.