We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Prediction of 35 Target Per- and Polyfluoroalkyl Substances (PFASs) in California Groundwater Using Multilabel Semisupervised Machine Learning
AI summary Read the abstract
Researchers built a machine learning pipeline trained on 25,000 groundwater observations across California to predict contamination levels of 35 individual PFAS compounds, achieving up to 100% accuracy and identifying contamination sources and soil characteristics as the strongest predictive factors.
Comprehensive monitoring of perfluoroalkyl and polyfluoroalkyl substances (PFASs) is challenging because of the high analytical cost and an increasing number of analytes. We developed a machine learning pipeline to understand environmental features influencing PFAS profiles in groundwater. By examining 23 public data sets (2016–2022) in California, we built a state-wide groundwater database (25,000 observations across 4200 wells) encompassing contamination sources, weather, air quality, soil, hydrology, and groundwater quality (PFASs and cocontaminants). We used supervised learning to prescreen total PFAS concentrations above 70 ng/L and multilabel semisupervised learning to predict 35 individual PFAS concentrations above 2 ng/L. Random forest with ADASYN oversampling performed the best for total PFASs (AUROC 99%). XGBoost with SMOTE oversampling achieved the AUROC of 73–100% for individual PFAS prediction. Contamination sources and soil variables contributed the most to accuracy. Individual PFASs were strongly correlated within each PFAS’s subfamily (i.e., short- vs long-chain PFCAs, sulfonamides). These associations improved prediction performance using classifier chains, which predicts a PFAS based on previously predicted species. We applied the model to reconstruct PFAS profiles in groundwater wells with missing data in previous years. Our approach can complement monitoring programs of environmental agencies to validate previous investigation results and prioritize sites for future PFAS sampling.
More Papers Like This
Predicted Potential for Aquatic Exposure Effects of Per- and Polyfluorinated Alkyl Substances (PFAS) in Pennsylvania’s Statewide Network of Streams
AI summary Read the abstract
This research uses predictive modeling to estimate aquatic exposure levels and ecological risk from per- and polyfluoroalkyl substances (PFAS), a class of persistent chemicals associated with plastic manufacturing and water contamination. The models identify specific PFAS compounds and environments at greatest risk of ecological harm. Findings can inform regulatory prioritization for PFAS monitoring and remediation in freshwater and marine systems.
Locating carcinogenic per- and poly-fluoroalkyl substances in Santa Clarita groundwater
AI summary Read the abstract
Using positive matrix factorization, this study identified domestic waste, fire extinguisher materials, and food packaging as three likely PFAS sources contaminating Santa Clarita's groundwater, though landfill leachate as the dominant source was not conclusively confirmed. PFAS and microplastics share origins in plastic packaging and consumer products, and co-contamination of drinking water supplies represents a compounding public health concern that underscores the need for comprehensive plastic pollution policy.
Interpretable machine learning models reveal the partnership of microplastics and perfluoroalkyl substances in sediments at a century scale
AI summary Read the abstract
Interpretable machine learning models applied to sediment core data from Taihu Lake revealed a strong spatial and temporal partnership between microplastics and perfluoroalkyl substances (PFASs), with MPs acting as co-transport vectors for PFASs. The study provided mechanistic insights into how these two classes of emerging contaminants interact in lacustrine sediments.
Identification of potentially contaminated areas of soil microplastic based on machine learning: A case study in Taihu Lake region, China
AI summary Read the abstract
Researchers applied machine learning models — including random forest and support vector regression — to predict the spatial distribution of soil microplastic pollution in China's Taihu Lake region, finding that soil texture, population density, and proximity to known plastic sources were the dominant drivers, with nearly half of urban soils showing serious contamination.
Application of machine learning in assessing spatial distribution patterns of soil microplastics: a case study of the Bang Pakong Watershed, Thailand
AI summary Read the abstract
Machine learning models were applied to predict spatial distribution patterns of microplastics in soils across a Thai watershed, identifying land use types and proximity to water bodies as key factors driving contamination levels.
Research digests by email
When a large batch of papers lands in the Atlas, we read through it and send a short write-up of what stood out.