This study collected 820 topsoil samples from cultivated lands across Ningxia, covering the Yellow River irrigation area, the central arid zone, and the southern mountainous region. The ordinary kriging were spatially interpolated to analyze As, Hg, Cd, Cr, and Pb heavy-metal pollution spatial patterns. Pollution was evaluated using the Nemerow and geoaccumulation (I_(geo)) indices, and sources quantified via Pearson correlations, PCA (Principal Component Analysis), and PMF (Positive Matrix Factorization). The results indicated that Hg and Cd posed the highest ecological risks. The overall mean concentrations (mg.kg⁻¹) of Hg, Cd, As, Pb, and Cr were 0.04, 0.27, 9.91,23.81, and 57.34, respectively. Compared with the background values, they were 1.90, 2.41, 0.83, 1.14, 2.74 times higher, respectively. Geospatially, regions with higher pollution probabilities for Cd, Cr, Pb, Hg, and As were concentrated in the northern and central parts of Ningxia, whereas the southern region exhibited lower pollution probabilities. pH significantly influenced the accumulation and spatial distribution of heavy metals in soil. Source apportionment identified three primary contributors: transportation and natural parent materials (As, Pb, Cr), industrial activities (Hg), and agricultural practices (Cd). Hg and Cd were identified as the key risk elements requiring prioritized management. These results enhance understanding of the pollution levers of heavy metals in Ningxia cultivated soils, and also provide foundation for developing more scientific and precise soil risk control policies, offering significant practical value for environmental risk management.

This study collected 820 topsoil samples from cultivated lands across Ningxia, covering the Yellow River irrigation area, the central arid zone, and the southern mountainous region. The concentrations of Arsenic (As), mercury (Hg), cadmium (Cd), chromium (Cr), and lead (Pb) were analyzed to evaluate the pollution levels and spatial distribution patterns of heavy metals, which were interpolated using Kriging. Pollution status was assessed via the Nemerow composite index and the geoaccumulation index (Igeo). Source identification and contribution quantification were performed using Pearson correlation analysis, principal component analysis (PCA), and positive matrix factorization (PMF). Results indicated that the cultivated soils in Ningxia are generally clean, with isolated instances of warning and light pollution. Ecological risk assessment revealed differential risk levels among elements, with Hg and Cd presenting the highest ecological risks. Geospatially, regions with higher pollution probabilities for Cd, Cr, Pb, Hg, and As were concentrated in the northern and central parts of Ningxia, whereas the southern region exhibited lower pollution probabilities.pH significantly influences the accumulation and spatial distribution of heavy metals in soil. Source apportionment identified three primary contributors: transportation and natural parent materials (As, Pb, Cr), industrial activities (Hg), and agricultural practices (Cd). Hg and Cd were identified as the key risk elements requiring prioritized management.

Accurate source identification is essential for the prevention and control of heavy metal(loid)s (HMs) contamination in farmland soils. Conventional source-apportionment approaches often rely on limited indicators and expert judgment, which can increase uncertainty in source interpretation. In this study, 800 topsoil samples collected from farmland in Ningxia, together with 24 environmental and anthropogenic variables, were used to develop element-specific machine learning models for Cd, Cr, Hg, Pb, and As. Five algorithms, including random forest (RF), Extra Trees (ET), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), and least absolute shrinkage and selection operator (LASSO-stacking), were compared, and Shapley additive explanations (SHAP) were used to interpret the variables statistically associated with the spatial variability of each metal. Positive matrix factorization (PMF) was further applied to identify potential source categories. Model performance varied among metals, indicating element-specific differences in predictability and controlling factors. SHAP analysis showed that precipitation, temperature, spatial coordinates (longitude/latitude), cropping intensity, and industrial-source fine particulate matter emissions were among the most important associated factors, although their relative importance differed across elements. PMF results suggested that 73.8% of Cd was associated with agricultural inputs, 87.6% of Hg with industrial atmospheric deposition, and 68.4% of Cr and 46.7% of As with natural sources, Pb showed relatively weak predictability, suggesting that its spatial variability may be influenced by unmeasured local or legacy inputs, while model-derived associations indicated possible links with spatial gradients, terrain-hydrological conditions, wind-related variables, and soil carrier properties. Overall, this study presents an integrated framework that combines machine learning-based associated-factor analysis with receptor-model source apportionment, providing a more nuanced understanding of HMs contamination in farmland soils and supporting targeted soil pollution prevention and control.

To characterize heavy metal accumulation, potential sources, and source-specific ecological risks in farmland soils of the Ningxia Yellow River Irrigation Area, 537 topsoil samples were analyzed for As, Hg, Cd, Pb, and Cr. positive matrix factorization (PMF) was coupled with the potential ecological risk index to quantify source-specific mass and ecological risk contributions. Cd and Hg showed the strongest enrichment relative to regional background values, with mean concentrations of 2.38 and 2.02 times the respective background values. PMF resolved four factors interpreted as an agricultural input-related source, a parent material-dominated natural source, an urban industrial- and combustion-related atmospheric deposition source, and a Yellow River alluvial–hydrological natural background source. The mean potential ecological risk index calculated from PMF reconstructed concentrations was 168.45, closely matching the observed value of 168.77. Source-specific mass and ecological risk contributions were clearly decoupled: the two natural source factors contributed 71.19% of the modeled heavy metal mass but only 24.21% of the ecological risk, whereas the two anthropogenic source factors contributed 28.81% of the mass but 75.80% of the risk. The atmospheric deposition and agricultural input-related sources contributed 47.27% and 28.53% of the ecological risk, respectively. Site-level bootstrap resampling and alternative allocation procedures retained the source risk ranking. These findings indicate that risk-based management should prioritize Hg-related atmospheric deposition and Cd-related agricultural inputs rather than total heavy metal mass alone.