Introduction: The Songnen Plain in Northeast China is considered as one of the world's three major saline-alkali land regions. However, the evaluation of the extent and degree of saline-alkali land has primarily been mainly obtained via remote sensing interpretation, leading to inconsistencies in data on salinization levels and the area of saline-alkali land.Methods: Random Forest-based modeling of total salt-exchangeable cation-chemical composition relationships facilitated high-precision spatial evaluation of soil salinization across Northeast China Plain.Results: The results show that: (1) Northeast China contains approximately 16.93 million hectares of salinized or alkalinized zonal soils, where pH and salinity levels demonstrate statistically significant positive correlations (p < 0.01) with concentrations of chloride (Cl), sodium (Na), magnesium (Mg), calcium (Ca), sulfur (S), and potassium (K). (2) Northeast China's Songliao Plain contains 6.92 million hectares of saline-alkali soils, exhibiting a distinct spatial gradient in soda salinization chemistry: from sulfate-dominated in the northeast, transitioning through chloride-sulfate composite zones, to carbonate and finally chloride-dominated types in the southwest. (3) Chloride and bicarbonate dominate the saline-alkali soils, primarily distributed in low plains ( 9.5) have expanded by 80,300 ha/year, accompanied by significant organic matter and nitrogen depletion.Discussion: Research indicates that integrated topography-hydrogeology-climate factors drive soil salinization, while consistent and stable farming practices can mitigate its progression.

Selenium is an important element for human health. Many studies have identified selenium deficiency in soil and water as an important factor in causing Keshan Disease (KD) in Northeast China. Previous studies have mainly focused on soil selenium content, staple food selenium content, and human selenium level, but there are few systematic studies on soil selenium’s existing forms and their migration from soil to crops and the human body. This paper focused on inferring the barrier factors in the migration of selenium from soil to crop and the human body and transformation of its compounds. It provides a reference basis for the etiological analysis, prevention, control, and elimination of KD. The study used 121 183 samples of topsoil (0−20 cm), 30 295 soil parent samples of selenium and other geochemical indices in northeast China, and crop seeds and human hair samples from the KD endemic area. The surface soil selenium was dominantly selenium-sufficient in Northeast China. However, the soil selenium levels were generally low. The average topsoil selenium in Northeast China was 0.20 mg/kg, significantly lower than the world’s average soil selenium content (0.4 mg/kg) and slightly lower than the Chinese average soil selenium content (0.24 mg/kg). Soil selenium mainly existed in strongly bounding by organic bound, with humic acid, and residue forms. The amount of selenium available to plants was sufficient in the selenium-sufficient and KD-endemic areas. However, the average selenium content of human hair was deficient, on average, with 0.16 mg/kg in KD endemic area. We assume that lower soil selenium content may be the basic factor influencing the biogeochemical deficiency of selenium in Northeast China. The sequestration of selenium by clay chemical constituents, such as iron and aluminum oxides and soil organic matter, especially in acidic soils, is another direct contributing factor to the low selenium content in biogeochemical food chain, which increases the risk of KD in the population.

Soil organic carbon (SOC) is a crucial indicator of soil quality and carbon cycling. While remote sensing and machine learning enable regional scale SOC prediction, most studies rely on vegetation indices (VIs) derived from bare-soil periods, potentially neglecting vegetation–soil interactions during crop growth. Given the bidirectional relationship between SOC and crop growth, we hypothesized that using crop-lush period VIs (VIs_lush) instead of bare-soil period VIs (VIs_bare) would increase the inversion accuracy. To test this hypothesis, we chose the cropland area in Liaoning Province as the study area and developed three modelling strategies (MS-1: VIs_lush + other features; MS-2: VIs_bare + other features; MS-3: without VIs) using Landsat 8 imagery, topographic and precipitation data, and ensemble learning models (XGBoost, RF, and AdaBoost), with SHapley Additive exPlanations (SHAP) analysis for variable interpretation. We found that 1) all models achieved their highest performance under MS-1, with XGBoost outperforming the others across all modelling strategies; 2) for XGBoost, MS-1 yielded the highest inversion accuracy (R² = 0.84, RMSE = 2.22 g·kg⁻¹, RPD = 2.49, and RPIQ = 3.25); compared with MS-2, MS-1 reduced the RMSE by 0.31 g·kg⁻¹, increased R² from 0.77 to 0.84, and reduced the RPD by 0.31 and the RPIQ by 0.40, and compared with MS-3, MS-1 reduced the RMSE by 0.41 g·kg⁻¹, increased R² from 0.79 to 0.84, and reduced the RPD by 0.39 and the RPIQ by 0.51; 2) SHAP analysis confirmed that VIs_lush contributed more than VIs_bare, supporting the rationale of using lush-period imagery; and 3) Liaoning Province exhibited distinct SOC spatial patterns (mean: 13.08 g·kg⁻¹), with values ranging from 2.19 g·kg⁻¹ (sandy central–western area) to 33.86 g·kg⁻¹ (eastern mountains/coast). This study demonstrates that integrating growth stage-specific VIs with ensemble learning can significantly enhance regional-scale SOC prediction.

Soil organic carbon (SOC) is a crucial indicator of soil quality and carbon cycling. While remote sensing and machine learning enable regional scale SOC prediction, most studies rely on vegetation indices (VIs) derived from bare-soil periods, potentially neglecting vegetation–soil interactions during crop growth. Given the bidirectional relationship between SOC and crop growth, we hypothesized that using crop-lush period VIs (VIs_lush) instead of bare-soil period VIs (VIs_bare) would increase the inversion accuracy. To test this hypothesis, we chose the cropland area in Liaoning Province as the study area and developed three modeling strategies (MS-1: VIs_lush + other features; MS-2: VIs_bare + other features; and MS-3: without VIs) using Landsat 8 imagery, topographic and precipitation data, and ensemble learning models (XGBoost, RF, and AdaBoost), with SHapley Additive exPlanations (SHAP) analysis for variable interpretation. We found that (1) all models achieved their highest performance under MS-1, with XGBoost outperforming the others across all modeling strategies; (2) for XGBoost, MS-1 yielded the highest inversion accuracy (R² = 0.84, RMSE = 2.22 g·kg⁻¹, RPD = 2.49, and RPIQ = 3.25); compared with MS-2, MS-1 reduced the RMSE by 0.31 g·kg⁻¹, increased R² from 0.77 to 0.84, and reduced the RPD by 0.31 and the RPIQ by 0.40, and compared with MS-3, MS-1 reduced the RMSE by 0.41 g·kg⁻¹, increased R² from 0.79 to 0.84, and reduced the RPD by 0.39 and the RPIQ by 0.51; (3) based on the SHAP analysis of the three modeling strategies, it is considered that precipitation, terrain and terrain analysis results are important indicators for SOC content inversion, and it is confirmed that VIs_lush contributed more than VIs_bare, supporting the rationale of using lush-period imagery; and (4) Liaoning Province exhibited distinct SOC spatial patterns (mean: 13.08 g·kg⁻¹), with values ranging from 2.19 g·kg⁻¹ (sandy central–western area) to 33.86 g·kg⁻¹ (eastern mountains/coast). This study demonstrates that integrating growth stage-specific VIs with ensemble learning can significantly enhance regional-scale SOC prediction.

Soil organic carbon (SOC) is central to terrestrial carbon cycling, soil health and sustainable land utilization. However, in Northeast China’s black soil region, the quantitative synergistic regulatory mechanisms of soil geochemical background, climate, and topography on SOC spatial heterogeneity remain poorly understood, limiting targeted sustainable soil carbon management. This study took the Hulan River Basin with a complete soil geochemical gradient as the study area. Based on eight major soil elements and pH data, we divided the basin into five geochemical zones via PCA and K-means clustering (cumulative variance contribution: 83.02%). SOC content differed significantly among zones, peaking in strongly acidic residual-slope zones and bottoming in alkaline saline–alkali alluvial zones. Geodetector results showed aridity (q = 0.57), mean annual temperature (q = 0.50), and geochemical zone (q = 0.46) dominated SOC differentiation. The non-linear interaction between geochemical zone and aridity exhibited the strongest explanatory power (q = 0.63), far exceeding individual effects, as revealed by the interaction detector. A dual regulatory mechanism was identified: iron–manganese oxides stabilize SOC via organo–mineral complexation, while high pH and base cations accelerate SOC decomposition by disrupting soil aggregates. This study clarifies the synergistic controls of pedogeochemistry, climate and topography on black soil SOC patterns, highlighting the underestimated dominant role of geochemical background. The findings support precise regional carbon stock assessment and differentiated carbon sequestration strategies.