Contamination of aquatic ecosystems by organic micropollutants (OMPs) like pharmaceuticals, personal care products, hormones, pesticides, and washing agents has raised concerns due to their bioaccumulative and toxic properties. Increasing OMP contamination in freshwater is a pressing public issue, spurring chemical regulations to assess risks and restrict or ban the most harmful compounds. Detecting OMPs in water requires advanced analytical techniques, often costly for monitoring programs. Machine Learning (ML) offers potential solutions: Predictive Modeling – training on datasets with known OMP concentrations and environmental parameters enables affordable prediction of OMP levels. Feature Selection – pinpointing key environmental variables predictive of OMP levels, optimizing data collection. Anomaly Detection – identifying outliers in new data for early detection. Since 2017, the Department of Aquatic Sciences and Assessment (IVM) at SLU has conducted monitoring studies for SEPA, collecting water samples from WWTP influents and effluents, as well as upstream and downstream surface waters in Sweden. This project aims to develop forecasting models to predict OMP levels in surface water, aiding in threat mitigation. Models will indicate potential exceedances of limit values, signaling the need for additional treatment at plants or upstream measures. The research impacts not just academia but supports the national goal of a 'non-toxic environment,' crucial for human and environmental health. By improving OMP monitoring in Swedish waters, we protect these ecosystems for current and future generations.
Project summary