Hydrogeochemical and sediment parameters improve predication accuracy of arsenic-prone groundwater in random forest machine-learning models

Recent advances in groundwater pollution research using machine learning from 2000 to 2023: A bibliometric analysis

Groundwater pollution has become a global challenge, posing significant threats to human health and ecological environments. Machine learning, with its superior ability to capture non-linear relationships in data, has shown significant potential in addressing groundwater pollution issues. This review presents a comprehensive bibliometric analysis of 1462 articles published between 2000 and 2023, offering an overview of the current state of research, analyzing development trends, and suggesting future directions. The analysis reveals a growing trend in publications over the 24-year period, with a sharp expansion since 2020. China, the USA, India, and Iran are identified as the leading contributors to publications and citations, with prominent institutions such as Jilin University, the United States Geological Survey, and the University of Tabriz. Moreover, keyword frequency analysis indicates that principal component analysis (PCA) is the most commonly used method, followed by artificial neural network (ANN) and hierarchical clustering analysis (HCA). The most studied groundwater pollutants include nitrate, arsenic, heavy metals, and fluoride. As machine learning has rapidly advanced, research focuses have evolved from fundamental tasks like hydrochemical evolution analysis, water quality index evaluation, and groundwater vulnerability assessments to more complex issues, such as pollutant concentration prediction, pollution risk assessment, and pollution source identification. Despite these advances, challenges related to data quality, data scarcity, model generalization, and interpretability remain. Future research should prioritize data sharing, improving model interpretability, broadening research horizons and advancing theory-guided machine learning. These will enhance our understanding of groundwater pollution mechanisms, and ultimately facilitate more effective pollution control and remediation strategies. In summary, this review provides valuable insights and suggestions for researchers and policymakers working in this critical field.

Assessing hydrogeochemical facies and Groundwater Quality Index in rapidly urbanizing coastal region: a GIS-based approach with machine learning for enhanced management

Groundwater is an essential freshwater source worldwide, but increasing pollution poses risks to its sustainability. This study applied a comprehensive approach to assess hydrogeochemical facies and groundwater quality in Odisha's large low-lying coastal regions. Analysis of 136 samples revealed that sodium (9.4%), potassium (40.8%), bicarbonate (2.1%), and chloride (2.1%) exceeded WHO limits. The Groundwater Quality Index (GQI) map classified 5.1% of samples as "excellent," 39.4% as "good," 31.3% as "poor," 13.8% as "very poor," and 10.2% as "unsuitable" for use. Additionally, the GQI values demonstrate a random spatial autocorrelation (- 0.06) likely due to diverse influences. The study identified the expansion of agricultural (43%) and built-up areas (13%) from the Land Use/Land Cover (LULC) map. Piper diagram and Gibbs plots suggest continued freshening, rock-water interaction, and seawater intrusion. Groundwater levels fall between 0 to 2 m below ground level (mbgl), primarily due to excessive groundwater extraction. The Sodium (Na+) vs. Chloride (Cl-) cross plot shows most samples align with the mixing line, with some deviations indicating multiple contamination sources. The strong correlation (> 0.90) between total dissolved salts (TDS), electrical conductivity (EC), Na+, and Cl- signals seawater intrusion, highlighting the complex interaction between human activities and natural processes. The proposed machine learning (ML) models like random forest (RF), artificial neural network (ANN), decision tree, and linear regression (LR) offer a reliable alternative to traditional GQI methods, addressing the challenges of extensive sampling and data management. Among these, RF exhibited the highest predictive accuracy (coefficient of correlation (R2) = 95%), surpassing ANN (R2 = 82%), decision tree (R2 = 81%), and LR (R2 = 67%) as the most effective model for GQI prediction. Potassium (K+) stands out as a key indicator of contamination. GQI, LULC map, and ML methods improve understanding of contamination sources and support systematic groundwater management.

Spatial distribution and hydrogeochemical processes of high iodine groundwater in the Hetao Basin, China

To understand the genesis and spatial distribution of high iodine groundwater in the Hetao Basin, 540 groundwater samples were analyzed for the chemistry and isotope. Total iodine concentrations in groundwater range from 1.32 to 2897 μg/L, with a mean value of 159.2 μg/L. The groundwater environment was mainly characterized by the weakly alkaline and reducing conditions, with the iodide as the main species of groundwater iodine. High iodine groundwater (I > 100 μg/L) was mainly distributed in shallow aquifers (< 30 m) of Hangjinhouqi near the Langshan Mountain and the discharge areas along the main drainage channels. The δ18O and δ2H values ranged from -12.09 ‰ to -3.99 ‰ and - 91.58 ‰ to -52.80 ‰, respectively, and the correlation between groundwater iodine and isotopes indicates the dominant role of evapotranspiration in the enrichment of iodine in the shallow groundwater with depth <30 m. It was further evidenced by the correlation between groundwater iodine and Cl/Br molar ratio, and significant contributions of climate factors identified from the random forest and XGBoost. Moreover, irrigation practices contribute to high iodine levels, with surface water used for irrigation containing up to 537.8 μg/L of iodine, which can be introduced into shallow aquifer directly. The iodine in irrigation water can be retained in the soil or shallow sediment, and later leach into groundwater under favorable conditions.

Utilizing machine learning to evaluate heavy metal pollution in the world's largest mangrove forest

The world's largest mangrove forest (Sundarbans) is facing an imminent threat from heavy metal pollution, posing grave ecological and human health risks. Developing an accurate predictive model for heavy metal content in this area has been challenging. In this study, we used machine learning techniques to model sediment pollution by heavy metals in this vital ecosystem. We collected 199 standardized sediment samples to predict the accumulation of eleven heavy metals using ten different machine learning algorithms. Among them, the extremely randomized tree model exhibited the best performance in predicting Fe (0.87), Cr (0.89), Zn (0.85), Ni (0.83), Cu (0.87), Co (0.62), As (0.68), and V (0.90), achieving notable R2 values. On the other hand, the random forest outperformed for predicting Cd (0.72) and Mn (0.91), whereas the decision tree model showed the best performance for Pb (0.73). The feature attribute analysis identified FeV, CrV, CuZn, CoMn, PbCd, and AsCd relationships resembled with correlation coefficients among them. Based on the established models, the prediction of the contamination factor of metals in sediments showed very high Cd contamination (CF ≥ 6). The Moran's I index for Cd, Cr, Pb, and As were 0.71, 0.81, 0.71, and 0.67, respectively, indicating strong positive spatial autocorrelation and suggesting clustering of similar contamination levels. Conclusively, this research provides a comprehensive framework for predicting heavy metal sediment pollution in the Sundarbans, identifying key areas needing urgent conservation. Our findings support the adoption of integrated management strategies and targeted remedial actions to mitigate the harmful effects of heavy metal contamination in this vital ecosystem.

Estimation of CO2-Brine Interfacial Tension Based on an Advanced Intelligent Algorithm Model: Application for Carbon Saline Aquifer Sequestration

The emission reduction of the main greenhouse gas, CO2, can be achieved via carbon capture, utilization, and storage (CCUS) technology. Geological carbon storage (GCS) projects, especially CO2 storage in deep saline aquifers, are the most promising methods for meeting the net zero emission goal. The safety and efficiency of CO2 saline aquifer storage are primarily controlled by structural and capillary trapping, which are significantly influenced by the interactions between fluid and solid phases in terms of the interfacial tension (IFT) between the injected CO2 and brine at the reservoir site. In this study, a model based on the random forest (RF) model and the Bayesian optimization (BO) algorithm was developed to estimate the IFT between the pure and impure gas-brine binary systems for application to CO2 saline aquifer sequestration. Then three heuristic algorithms were applied to validate the accuracy and efficiency of the established model. The results of this study indicate that among the four mixed models, the Bayesian optimized random forest model fits the experimental data with the smallest root-mean-square error (RMSE = 1.7705) and mean absolute percentage error (MAPE = 2.0687%) and a high coefficient of determination (R2 = 0.9729). Then the IFT values predicted via this model were used as an input parameter to estimate the CO2 sequestration capacity of saline aquifers at different depths in the Tarim Basin of Xinjiang, China. The burial depth had a limited influence on the CO2 storage capacity.

Optimizing cropland use to reduce groundwater arsenic hazards in a naturally arsenic-enriched grain-producing region

In the Hetao Basin, a grain-producing region plagued by naturally occurring arsenic (As) pollution, understanding the role of agricultural cultivation activities in mobilizing As in groundwater is worthwhile. Here we investigated the impact of cropland use characteristics on groundwater As hazards using a model that combines Random Forest (RF) classification with SHapley Additive exPlanation (SHAP). The analysis incorporated eight cropland use characteristics and three natural factors across 1258 groundwater samples as independent variables. Additionally, an optimized cropland use strategy to mitigate groundwater As hazards was proposed. The results revealed that crop cultivation area, especially within a 2500m-radius buffer around sampling points, most significantly influenced the probability of groundwater As concentrations exceeding an irrigation safety threshold of 50 μg/L, achieving an AUC of 0.86 for this prediction. The relative importance of crop areas on As hazards were as follows: sunflower > melon > wheat > maize. Specifically, a high proportion of sunflower area (>30%), particularly in regions with longer cropland irrigation history, tended to elevate groundwater As hazards. Conversely, its negative driving force on groundwater As hazards was more pronounced with the increase in the proportion of wheat area (>5%), in contrast to other crops. Transitioning from sunflower to wheat or melon cultivation in the northeast of the Hetao Basin may contribute to lower groundwater As hazards. This study provides a scientific foundation for balancing food production with environmental safety and public health considerations.

Groundwater salinity modeling and mapping using machine learning approaches: a case study in Sidi Okba region, Algeria

The groundwater salinization process complexity and the lack of data on its controlling factors are the main challenges for accurate predictions and mapping of aquifer salinity. For this purpose, effective machine learning (ML) methodologies are employed for effective modeling and mapping of groundwater salinity (GWS) in the Mio-Pliocene aquifer in the Sidi Okba region, Algeria, based on limited dataset of electrical conductivity (EC) measurements and readily available digital elevation model (DEM) derivatives. The dataset was randomly split into training (70%) and testing (30%) sets, and three wrapper selection methods, recursive feature elimination (RFE), forward feature selection (FFS), and backward feature selection (BFS) are applied to train the data. The resulting combinations are used as inputs for five ML models, namely random forest (RF), hybrid neuro-fuzzy inference system (HyFIS), K-nearest neighbors (KNN), cubist regression model (CRM), and support vector machine (SVM). The best-performing model is identified and applied to predict and map GWS across the entire study area. It is highlighted that the applied methods yield input variation combinations as critical factors that are often overlocked by many researchers, which substantially impacts the models' accuracy. Among different alternatives the RF model emerged as the most effective for predicting and mapping GWS in the study area, which led to the high performance in both the training (RMSE = 1.016, R = 0.854, and MAE = 0.759) and testing (RMSE = 1.069, R = 0.831, and MAE = 0.921) phases. The generated digital map highlighted the alarming situation regarding excessive GWS levels in the study area, particularly in zones of low elevations and far from the Foum Elgherza dam and Elbiraz wadi. Overall, this study represents a significant advancement over previous approaches, offering enhanced predictive performance for GWS with the minimum number of input variables.

Hazards and influence factors of arsenic in the upper pleistocene aquifer, Hetao region, using machine learning modeling

The Hetao region is one of the regions with the most serious problem of the greatest measured arsenic concentrations in China. The enrichment of arsenic in groundwater may poses a great risk to the health of local residents. A comprehensive understanding of the groundwater quality, spatial distribution characteristics and hazard of the high arsenic in groundwater is indispensable for the sustainable utilization of groundwater resources and resident health. This study selected six environmental factors, climate, human activity, sedimentary environment, hydrogeology, soil, and others, as the independent input variables to the model, compared three machine learning algorithms (support vector machine, extreme gradient boosting, and random forest), and mapped unsafe arsenic to estimate the population that may be exposed to unhealthy conditions in the Hetao region. The results show that nearly half the number of the 605 sampling wells for arsenic exceeded the WHO provisional guide value for drinking water, the water chemistry of groundwater are mainly Na-HCO3-Cl or Na-Mg-HCO3-Cl type water, and the groundwater with excessive arsenic concentration is mainly concentrated in the ancient stream channel influence zone and the Yellow River crevasse splay. The results of factor importance explanation revealed that the sedimentary environment was the key factor affecting the primary high arsenic groundwater concentration, followed by climate and human activities. The random forest algorithm produced the probability distribution of high arsenic groundwater that is consistent with the observed results. The estimated area of groundwater with excessive arsenic reached 38.81 %. An estimated 940,000 people could be exposed to high arsenic in groundwater.