In recent years, excessive accumulations of iron (Fe), manganese (Mn), and nitrogen (N) have been observed in the groundwater of agricultural regions, particularly in flood irrigation areas. Nevertheless, the causes of this phenomenon and the associated hydrobiogeochemical processes remain elusive. This study demonstrated that redox fluctuations instigated by flood irrigation triggered a synergistic interaction between the N cycles and the activation of Fe and Mn oxides, thereby resulting in elevated concentrations of Fe, Mn, and N simultaneously. Static experiments revealed that the properties of the topsoil exerted a profound influence on the N induced release of Fe and Mn. The black soil (TFe: 1.5-2.3 times, Mn(II): 1.1-1.5 times, nitrate: 1.3-1.4 times) had greater release potential than meadow and dark brown soils due to higher electron donors/acceptors and substrates. Dynamic column experiments further elucidated that the wet-dry cycles induced by agricultural cultivation regulated the release process through the formation of zonal redox gradients and the structuring of microbial community. Organic nitrogen mineralization, chemolithotrophic nitrification, and Feammox/Mnammox were identified as the primary mechanisms responsible for the reductive dissolution of Fe-Mn oxides. On the other hand, autotrophic denitrification, with nitrate serving as the electron acceptor, constituted the main process for the reoxidation of Fe and Mn. Furthermore, the agricultural activities exerted a significant impact on the nitrate attenuation process, ultimately resulting in the recurrence of TFe (black soil: 1.5-6.3 times) and nitrate (black soil: 1.4-1.6 times) pollution during the phase after harvesting of rice (days 40-45) in saturated zone. The findings of this study not only deepened the understanding of the intricate interactions and coupled cycles between primary geochemical compositions and anthropogenic pollutants, but also provided a scientific foundation for the effective management and prevention of groundwater pollution stemming from agricultural cultivation processes.

Gas station sites pose potential risks of soil and groundwater contamination, which not only threatens public health and property but may also damage the assets and reputation of businesses and government entities. Given the complex nature of soil and groundwater contamination at gas station sites, this study utilizes field data from basic and environmental information, maintenance information for tank and pipeline monitoring, and environmental monitoring to develop machine learning models for predicting potential contamination risks and evaluating high-impact risk factors. The research employs three machine learning models: XGBoost, LightGBM, and Random Forest (RF). To compare the performance of these models in predicting soil and groundwater contamination, multiple performance metrics were utilized, including Receiver Operating Characteristic (ROC) curves, Precision-Recall graphs, and Confusion Matrix (CM). The Confusion Matrix analysis revealed the following results: accuracy of 85.1-87.4 %, precision of 86.6-88.3 %, recall of 83.0-87.2 %, and F1 score of 84.8-87.8 %. Performance ranking across all metrics consistently showed: XGBoost > LightGBM > RF. The area under the ROC curve and precision-recall curve for the three models were 0.95 (XGBoost), 0.94 (LightGBM), and 0.93 (RF), respectively. While all three machine learning approaches demonstrated satisfactory predictive capabilities, the XGBoost model exhibited optimal performance across all evaluation metrics. This research demonstrates that properly trained machine learning models can serve as effective tools for environmental risk assessment and management. These findings have significant implications for decision-makers in environmental protection, enabling more accurate prediction and control of contamination risks, thereby enhancing the preservation of ecological systems, public health, and property security.