A 40-year dataset of soil salinity dynamics (1985-2024) at 100 m resolution in the Western Songnen Plain, China

Soil salinization leads to land degradation, reduced agricultural productivity, and heightened food security risks. Accurate assessment of saline soil distribution and severity is critical for sustainable land management. However, challenges such as broad spatial extent, high heterogeneity, and limited field observations hinder mapping accuracy. Existing datasets in China show large discrepancies in salinized area estimates due to coarse spatial and temporal resolution. In this study, we classified soil salinity degree across the western Songnen Plain from 1985 to 2024 using field surveys, remote sensing imagery, and machine learning algorithms, achieving high accuracy (overall accuracy = 0.893, Kappa = 0.782). A regional soil EC prediction model (R² = 0.467) was developed using 942 in situ samples and remotely sensed indicators, accounting for soil moisture effects. This model produced annual, 100 m resolution maps from 1985 to 2024, with only 2.78% deviation from the second National Land Survey. The resulting high-resolution dataset reveals the spatiotemporal dynamics of soil salinity and supports improved monitoring and management to address environmental sustainability and food security.