Portable X-ray fluorescence (pXRF) spectrometers provide simple, rapid, nondestructive, and cost-effective analysis of the metal contents in soils. The current method for improving pXRF measurement accuracy is soil sample preparation, which inevitably consumes significant amounts of time. To eliminate the influence of sample preparation on PXRF measurements, this study evaluates the performance of pXRF measurements in the prediction of eight heavy metals’ contents through machine learning algorithm linear regression (LR) and multivariate adaptive regression spline (MARS) models. Soil samples were collected from five industrial sites and separated into high-value and low-value datasets with pXRF measurements above or below the background values. The results showed that for Cu and Cr, the MARS models were better than the LR models at prediction (the MARS-R² values were 0.88 and 0.78; the MARS-RPD values were 2.89 and 2.11). For the pXRF low-value dataset, the multivariate MARS models improved the pXRF measurement accuracy, with the R² values improved from 0.032 to 0.39 and the RPD values increased by 0.02 to 0.37. For the pXRF high-value dataset, the univariate MARS models predicted the content of Cu and Cr with less calculation. Our study reveals that machine learning methods can better predict the Cu and Cr of large samples from multiple contaminated sites.

Chlorinated aliphatic hydrocarbons (CAHs) are widely used in agriculture and industries and have become one of the most common groundwater contaminations. With the excellent performance of the deep learning method in predicting, LSTM and XGBoost were used to forecast dichloroethene (DCE) concentrations in a pesticide-contaminated site undergoing natural attenuation. The input variables included BTEX, vinyl chloride (VC), and five water quality indicators. In this study, the predictive performances of long short-term memory (LSTM) and extreme gradient boosting (XGBoost) were compared, and the influences of variables on models’ performances were evaluated. The results indicated XGBoost was more likely to capture DCE variation and was robust in high values, while the LSTM model presented better accuracy for all wells. The well with higher DCE concentrations would lower the model’s accuracy, and its influence was more evident in XGBoost than LSTM. The explanation of the SHapley Additive exPlanations (SHAP) value of each variable indicated high consistency with the rules of biodegradation in the real environment. LSTM and XGBoost could predict DCE concentrations through only using water quality variables, and LSTM performed better than XGBoost.

The accumulation of nitrogen in groundwater in the industrial plots, especially the high ammonium, can result in a serious threat to the groundwater system in the urban area. This study monitored the dissolved inorganic nitrogen (DIN) of the polluted groundwater four times in one year in a retired nitrogenous fertilizer plant site with a production history of nearly 40 years, to analyze the spatial-temporal characteristics of DIN species (NH₄⁺-N, NO₃⁻-N, and NO₂⁻-N) and the effects of groundwater environment on their transfer and transformation. The results showed that NH₄⁺-N (<0.025 to 1310 mg/L) was the main DIN species (61.38–76.80%) with low mobility, whereas the concentration of NO₃⁻-N and NO₂⁻-N was 0.15–146 mg/L and <0.001–12.4 mg/L, accounting for 22.34–36.07% and 0.53–2.83% of total DIN, respectively. The concentration and proportion of NO₃⁻-N and NO₂⁻-N showed an upward trend with time, posing a threat to the safety of surrounding groundwater, and their high spatial-temporal variation was related to the morphological transformation and the transport. In the wet season, the pH and redox condition benefited the nitrification, and NO₃⁻-N easily migrated from the deep soil solution to groundwater, hence the NO₃⁻-N can be accumulated. Therefore, the analysis of species and behaviors of DIN in shallow groundwater is indispensable for environmental risk assessment.

Soil heavy metals (HMs) pollution from mining activities poses a significant threat to environmental health and has attracted widespread attention. This study integrated eco-health risk assessment and identified driving factors for risks of HMs in soils at an open-pit lead-zinc mining area. Results indicated that Pb, As and Cu were the major contaminants, with 46.43%, 32.14% and 25.00% of soil samples exhibited moderate to severe pollution levels. Distinct from the pollution level, Cd total contents and effective available contents showed the highest ecological risk, with an average Er value of 180.06 (heavily polluted) and mean RAC value of 82.01(extreme pollution), respectively. Through the probabilistic health risk assessment, As was identified as the key pollutant posing health risks, contributing 98.6% and 97.8% of the Carcinogenic risks and 68.5% and 76.7% of the Non-carcinogenic risks for adults and children, respectively. Cd and As were identified as the primary pollutants posing potential ecological and human health risks. Redundancy and permutation feature importance analysis further demonstrated that electrical conductivity (EC) was a key predictor strongly associated with the ecological risk and health risk. This study distinguished the target priority pollutant between the ecological and health risks and identified the driving factors for risks, providing valuable insights into the pollution control and targeted soil management in mining areas.

Humic substances (HSs) possess abundant oxygen-containing functional groups that theoretically confer exceptional complexation capacity for heavy metal (HM) remediation; however, their speciation-dependent extractive behaviors across multi-metal contaminated soils remain insufficiently quantified. This study systematically evaluated the washing performance of six humic-based agents derived from weathered coal and lignite on soils contaminated with representative cadmium and arsenic species. Batch sequential washing experiments demonstrated that a four-cycle extraction at a liquid-to-solid ratio of 30:1 achieved high average cumulative removal efficiencies of 92.9% for Cd and 83.0% for As, with weathered coal-derived humic acid (HAW) exhibiting the highest removal efficiency. Microstructural and spectroscopic characterizations indicated that abundant oxygen-containing functional groups and highly condensed aromatic structures served as the core drivers for HM complexation. Furthermore, the XGBoost-SHAP model accurately predicted Cd leaching efficiency, identifying washing times as the most influential variable, followed by contaminant speciation, leaching agent type, and L/S ratio. The substantially lower performance for As suggests that arsenic leaching is governed by additional physicochemical factors beyond the four operational variables, highlighting the need for speciation-related descriptors in future modeling efforts. Overall, this work demonstrates the technical feasibility of converting low-rank coal resources into high-value, sustainable soil washing agents, offering robust data-driven insights for the targeted design of site remediation strategies.