Integrated analytical framework for soil salinization assessment in the Yellow River Delta coastline: Spatiotemporal dynamics, future trends, and driving mechanisms (2003–2023)

Coastal salinization threatens arable land where marine influence, shallow groundwater, and cultivation interact. We develop an integrated framework that unifies spatiotemporal trend detection (Sen’s slope–Mann–Kendall), future trend persistence (Hurst exponent), and driver attribution with interactions (GeoDetector) to assess soil salinity content (SSC) in the Yellow River Delta from 2003 to 2023. SSC was mapped annually from Landsat using a random-forest model (validation RMSE 1.55 g kg⁻¹; RPD 1.60). Regional mean SSC declined from 2.46 to 2.20 g kg⁻¹(−10.6 %). Significant improvement covers 46.9 % of land, slight improvement 16.5 %, slight degradation 28.7 %, and significant degradation 3.0 %, indicating predominant improvement with localized relapse. Persistence analysis shows 46.9 % of the area is likely to sustain improvement, 16.5 % has potential to improve, about 29 % is at risk of renewed salinization, and 3 % exhibits persistent degradation; high-risk tracts cluster in southern Guangrao, eastern Wudi, western Zhanhua, and along the Yellow River. Driver attribution ranks the fallow-land ratio as the leading anthropogenic control, and distance to the sea, elevation, evapotranspiration, and silt content as dominant natural drivers; most two-factor combinations show nonlinear enhancement. We conclude that stabilizing cultivation and maintaining drainage underpin durable gains, while risk belts require targeted leaching, drain rehabilitation, and salt-tolerant rotations. By overcoming the separation between change diagnosis, future risk, and governance, this framework couples “where” and “how,” and supports management under different salinization conditions, future trajectories, and dominant drivers in coastal deltas.