Groundwater-anomaly-driven reorganization and threshold response of soil salinity profiles in the Shandong coastal low plain

Soil salinization in the coastal lowland plain of Shandong is not merely surface salt accumulation, but a profile-scale redistribution of water and salts driven by shallow groundwater, strong evaporation, and land-use disturbance. However, how groundwater anomalies reorganize salinity across soil layers and whether these responses vary nonlinearly among land-use types remain poorly quantified. To address this gap, this study integrated monthly groundwater records, soil salinity data from the 0–5, 5–15, and 15–30 cm layers, remote sensing products, meteorological–hydrological variables, and land-use data from 2010 to 2022, and employed hydrogeological process indices, spatially blocked machine learning, SHAP interpretation, and counterfactual segmented regression to identify soil salinity profile reorganization patterns and groundwater-depth response breakpoints. The model explained salinity variability well, with an R² of 0.72. Results showed a shift from conventional surface enrichment to multi-pathway profile reorganization, including surface accumulation, mid-layer enrichment, and downward migration. Over the 13-year period, surface salt accumulation in cropland increased from 22% to 50%, while downward migration in forestland rose from 35% to 46%, revealing contrasting profile-reorganization pathways. Groundwater-depth anomaly showed the largest model contribution. The overall breakpoint was 0.42 m, beyond which the salinity response slope increased from 0.041 to 0.183. Land-use-specific thresholds were 0.39, 0.31, 0.36, and 0.67 m for cropland, built-up-adjacent soils, bare land, and forestland. These thresholds identify salinization-sensitive groundwater conditions and provide quantitative guidance for groundwater monitoring and land-use-specific irrigation–drainage management.