Groundwater-related land subsidence is widespread in Iran’s alluvial aquifers, yet the spatial correspondence between groundwater use, land-use/land-cover (LULC) patterns, and surface deformation remains poorly constrained. This study integrated annual Sentinel-1 SBAS-InSAR line-of-sight (LOS) deformation for 2015–2024, Random Forest LULC maps for 2010, 2015, 2020, and 2024, a provincial pumping-well inventory, and temporally resolved groundwater-monitoring records across the Tehran–Karaj, Hashtgerd, and Eshtehard plains of Alborz Province. Direct LULC–deformation comparisons were restricted to exact same-year pairs for 2015, 2020, and 2024, and spatial dependence was addressed using variograms, distance thinning, and 14-km block-level inference. Agricultural land declined by 98.99 km² (14.80%) between 2010 and 2024. Within the valid long-term InSAR footprint, 636.14 km² (19.24%) met the operational active-subsidence criterion of ≤ −20 mm yr⁻¹. Agricultural land had the most-negative mean and median annual LOS deformation in all three matched years, with Kruskal–Wallis ε² values of 0.378, 0.222, and 0.273. This ranking persisted in 42 of 45 spatial-thinning configurations, while the agriculture–bare-land contrast remained supported at the 14-km block scale in all three years. Static well data showed a further spatial mismatch: Tehran–Karaj contained the largest recorded pumping capacity, whereas Hashtgerd exhibited substantially more-negative long-term deformation. The recorded-discharge and subsidence-intensity centroids were separated by 15.16 km, and this distance varied only from 15.00 to 15.28 km across tested LOS thresholds. In contrast, a 23-site temporal monitoring panel showed no robust contemporaneous or 1–2-year lagged association between monitored pumping intensity and annual LOS deformation. The results indicate that groundwater-use environments are persistently associated with stronger subsidence, but recorded pumping capacity alone does not explain its spatial distribution. Hydrogeological heterogeneity and groundwater-head history are therefore critical for interpreting and managing subsidence across the three plains.

Groundwater-related land subsidence is widespread in Iran’s alluvial aquifers, yet the spatial correspondence between groundwater use, land-use/land-cover (LULC) patterns, and surface deformation remains poorly constrained. This study integrated annual Sentinel-1 SBAS-InSAR line-of-sight (LOS) deformation for 2015–2024, Random Forest LULC maps for 2010, 2015, 2020, and 2024, a provincial pumping-well inventory, and temporally resolved groundwater-monitoring records across the Tehran–Karaj, Hashtgerd, and Eshtehard plains of Alborz Province. Direct LULC–deformation comparisons were restricted to exact same-year pairs for 2015, 2020, and 2024, and spatial dependence was addressed using variograms, distance thinning, and 14-km block-level inference. Agricultural land declined by 98.99 km² (14.80%) between 2010 and 2024. Within the valid long-term InSAR footprint, 636.14 km² (19.24%) met the operational active-subsidence criterion of ≤ −20 mm yr⁻¹. Agricultural land had the most-negative mean and median annual LOS deformation in all three matched years, with Kruskal–Wallis ε² values of 0.378, 0.222, and 0.273. This ranking persisted in 42 of 45 spatial-thinning configurations, while the agriculture–bare-land contrast remained supported at the 14-km block scale in all three years. Static well data showed a further spatial mismatch: Tehran–Karaj contained the largest recorded pumping capacity, whereas Hashtgerd exhibited substantially more-negative long-term deformation. The recorded-discharge and subsidence-intensity centroids were separated by 15.16 km, and this distance varied only from 15.00 to 15.28 km across tested LOS thresholds. In contrast, a 23-site temporal monitoring panel showed no robust contemporaneous or 1–2-year lagged association between monitored pumping intensity and annual LOS deformation. The results indicate that groundwater-use environments are persistently associated with stronger subsidence, but recorded pumping capacity alone does not explain its spatial distribution. Hydrogeological heterogeneity and groundwater-head history are therefore critical for interpreting and managing subsidence across the three plains.

Groundwater-related land subsidence is widespread in Iran’s alluvial aquifers, yet the spatial correspondence between groundwater use, land-use/land-cover (LULC) patterns, and surface deformation remains poorly constrained. This study integrated annual Sentinel-1 SBAS-InSAR line-of-sight (LOS) deformation for 2015–2024, Random Forest LULC maps for 2010, 2015, 2020, and 2024, a provincial pumping-well inventory, and temporally resolved groundwater-monitoring records across the Tehran–Karaj, Hashtgerd, and Eshtehard plains of Alborz Province. Direct LULC–deformation comparisons were restricted to exact same-year pairs for 2015, 2020, and 2024, and spatial dependence was addressed using variograms, distance thinning, and 14-km block-level inference. Agricultural land declined by 98.99 km² (14.80%) between 2010 and 2024. Within the valid long-term InSAR footprint, 636.14 km² (19.24%) met the operational active-subsidence criterion of ≤ −20 mm yr⁻¹. Agricultural land had the most-negative mean and median annual LOS deformation in all three matched years, with Kruskal–Wallis ε² values of 0.378, 0.222, and 0.273. This ranking persisted in 42 of 45 spatial-thinning configurations, while the agriculture–bare-land contrast remained supported at the 14-km block scale in all three years. Static well data showed a further spatial mismatch: Tehran–Karaj contained the largest recorded pumping capacity, whereas Hashtgerd exhibited substantially more-negative long-term deformation. The recorded-discharge and subsidence-intensity centroids were separated by 15.16 km, and this distance varied only from 15.00 to 15.28 km across tested LOS thresholds. In contrast, a 23-site temporal monitoring panel showed no robust contemporaneous or 1–2-year lagged association between monitored pumping intensity and annual LOS deformation. The results indicate that groundwater-use environments are persistently associated with stronger subsidence, but recorded pumping capacity alone does not explain its spatial distribution. Hydrogeological heterogeneity and groundwater-head history are therefore critical for interpreting and managing subsidence across the three plains.Groundwater-related land subsidence is widespread in Iran’s alluvial aquifers, yet the spatial correspondence between groundwater use, land-use/land-cover (LULC) patterns, and surface deformation remains poorly constrained. This study integrated annual Sentinel-1 SBAS-InSAR line-of-sight (LOS) deformation for 2015–2024, Random Forest LULC maps for 2010, 2015, 2020, and 2024, a provincial pumping-well inventory, and temporally resolved groundwater-monitoring records across the Tehran–Karaj, Hashtgerd, and Eshtehard plains of Alborz Province. Direct LULC–deformation comparisons were restricted to exact same-year pairs for 2015, 2020, and 2024, and spatial dependence was addressed using variograms, distance thinning, and 14-km block-level inference. Agricultural land declined by 98.99 km² (14.80%) between 2010 and 2024. Within the valid long-term InSAR footprint, 636.14 km² (19.24%) met the operational active-subsidence criterion of ≤ −20 mm yr⁻¹. Agricultural land had the most-negative mean and median annual LOS deformation in all three matched years, with Kruskal–Wallis ε² values of 0.378, 0.222, and 0.273. This ranking persisted in 42 of 45 spatial-thinning configurations, while the agriculture–bare-land contrast remained supported at the 14-km block scale in all three years. Static well data showed a further spatial mismatch: Tehran–Karaj contained the largest recorded pumping capacity, whereas Hashtgerd exhibited substantially more-negative long-term deformation. The recorded-discharge and subsidence-intensity centroids were separated by 15.16 km, and this distance varied only from 15.00 to 15.28 km across tested LOS thresholds. In contrast, a 23-site temporal monitoring panel showed no robust contemporaneous or 1–2-year lagged association between monitored pumping intensity and annual LOS deformation. The results indicate that groundwater-use environments are persistently associated with stronger subsidence, but recorded pumping capacity alone does not explain its spatial distribution. Hydrogeological heterogeneity and groundwater-head history are therefore critical for interpreting and managing subsidence across the three plains.

Groundwater-related land subsidence is widespread in Iran’s alluvial aquifers, yet the spatial correspondence between groundwater use, land-use/land-cover (LULC) patterns, and surface deformation remains poorly constrained. This study integrated annual Sentinel-1 SBAS-InSAR line-of-sight (LOS) deformation for 2015–2024, Random Forest LULC maps for 2010, 2015, 2020, and 2024, a provincial pumping-well inventory, and temporally resolved groundwater-monitoring records across the Tehran–Karaj, Hashtgerd, and Eshtehard plains of Alborz Province. Direct LULC–deformation comparisons were restricted to exact same-year pairs for 2015, 2020, and 2024, and spatial dependence was addressed using variograms, distance thinning, and 14-km block-level inference. Agricultural land declined by 98.99 km² (14.80%) between 2010 and 2024. Within the valid long-term InSAR footprint, 636.14 km² (19.24%) met the operational active-subsidence criterion of ≤ −20 mm yr⁻¹. Agricultural land had the most-negative mean and median annual LOS deformation in all three matched years, with Kruskal–Wallis ε² values of 0.378, 0.222, and 0.273. This ranking persisted in 42 of 45 spatial-thinning configurations, while the agriculture–bare-land contrast remained supported at the 14-km block scale in all three years. Static well data showed a further spatial mismatch: Tehran–Karaj contained the largest recorded pumping capacity, whereas Hashtgerd exhibited substantially more-negative long-term deformation. The recorded-discharge and subsidence-intensity centroids were separated by 15.16 km, and this distance varied only from 15.00 to 15.28 km across tested LOS thresholds. In contrast, a 23-site temporal monitoring panel showed no robust contemporaneous or 1–2-year lagged association between monitored pumping intensity and annual LOS deformation. The results indicate that groundwater-use environments are persistently associated with stronger subsidence, but recorded pumping capacity alone does not explain its spatial distribution. Hydrogeological heterogeneity and groundwater-head history are therefore critical for interpreting and managing subsidence across the three plains.

Groundwater-related land subsidence is widespread in Iran’s alluvial aquifers, yet the spatial correspondence between groundwater use, land-use/land-cover (LULC) patterns, and surface deformation remains poorly constrained. This study integrated annual Sentinel-1 SBAS-InSAR line-of-sight (LOS) deformation for 2015–2024, Random Forest LULC maps for 2010, 2015, 2020, and 2024, a provincial pumping-well inventory, and temporally resolved groundwater-monitoring records across the Tehran–Karaj, Hashtgerd, and Eshtehard plains of Alborz Province. Direct LULC–deformation comparisons were restricted to exact same-year pairs for 2015, 2020, and 2024, and spatial dependence was addressed using variograms, distance thinning, and 14-km block-level inference. Agricultural land declined by 98.99 km² (14.80%) between 2010 and 2024. Within the valid long-term InSAR footprint, 636.14 km² (19.24%) met the operational active-subsidence criterion of ≤ −20 mm yr⁻¹. Agricultural land had the most-negative mean and median annual LOS deformation in all three matched years, with Kruskal–Wallis ε² values of 0.378, 0.222, and 0.273. This ranking persisted in 42 of 45 spatial-thinning configurations, while the agriculture–bare-land contrast remained supported at the 14-km block scale in all three years. Static well data showed a further spatial mismatch: Tehran–Karaj contained the largest recorded pumping capacity, whereas Hashtgerd exhibited substantially more-negative long-term deformation. The recorded-discharge and subsidence-intensity centroids were separated by 15.16 km, and this distance varied only from 15.00 to 15.28 km across tested LOS thresholds. In contrast, a 23-site temporal monitoring panel showed no robust contemporaneous or 1–2-year lagged association between monitored pumping intensity and annual LOS deformation. The results indicate that groundwater-use environments are persistently associated with stronger subsidence, but recorded pumping capacity alone does not explain its spatial distribution. Hydrogeological heterogeneity and groundwater-head history are therefore critical for interpreting and managing subsidence across the three plains.