Study finds emerging contaminants follow different sediment-water rules

Researchers in China say three major classes of emerging contaminants — PFASs, antibiotics and endocrine-disrupting chemicals — do not partition between sediment and water in the same way. The findings, published Aug. 14 in Environmental Science and Ecotechnology, could improve basin-scale monitoring and risk management in rivers.

Why it matters: - Emerging contaminants can settle into river sediments or stay mobile in water, and that behavior shapes how long pollutants persist and where they pose the biggest risk. - Better prediction of sediment-water partitioning could help agencies target monitoring, prioritize hotspots and improve wastewater and discharge planning.

What happened: - Researchers at Sun Yat-sen University in Guangzhou, China, built a cross-scale predictive framework for three major contaminant classes: PFASs, antibiotics and endocrine-disrupting chemicals. - The study was published Aug. 14, 2026, in Environmental Science and Ecotechnology and is identified by DOI 10.1016/j.ese.2026.100751. - The team found that the three classes follow distinct mechanisms when partitioning between sediment and water.

The details: - The researchers compiled 5,085 paired sediment-water records from 1,093 sampling sites across China's seven major river basins. - The team used a multi-branch multi-head attention machine learning model that separately processed molecular descriptors, sediment-water properties and basin characteristics. - The model achieved R² values of 0.76 for PFASs, 0.92 for antibiotics and 0.89 for endocrine-disrupting chemicals. - Those results outperformed conventional artificial neural networks. - Molecular dynamics simulations supported the class-specific mechanisms at the molecular interface. - PFAS compounds formed sodium-bridged complexes with mineral surfaces. - Antibiotics showed stronger affinity for organic matter. - The framework also produced spatial maps showing high-accumulation and high-mobility zones. - The team used the framework to generate monthly log Kd projections across the Greater Bay Area, a densely urbanized region.

Between the lines: - Antibiotics appear to be the most straightforward class to predict because molecular structure alone explains much of their behavior. - PFASs are more sensitive to environmental conditions such as temperature and pH, so local water chemistry matters more. - EDCs sit in the middle, with influences from molecular properties, water conditions and basin-scale urbanization. - A one-size-fits-all model would miss the different drivers behind each contaminant class.

What's next: - The framework could help identify where contaminants are likely to accumulate in sediments versus remain mobile in the water column. - The resulting maps may help prioritize monitoring sites, optimize discharge schedules and support adaptive wastewater treatment strategies. - The study was funded by the National Natural Science Foundation of China, the General Program of the National Science Foundation of Guangdong Province and the National Key Research and Development Program of China.