Abstract
Per- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants with complex ecosystem dynamics, yet the mechanisms governing their environmental fate remain poorly understood. Assessing the relative importance of different exposure routes and the influence of environmental conditions on PFAS distribution is crucial for evaluating ecological and human health risks. Traditional bioaccumulation models based on lipid partitioning are unsuitable for PFAS, and comprehensive physicochemical property data for many PFAS is still unavailable. Testing every PFAS at all levels of an ecosystem is impossible, and in the absence of crucial data, individual compound-based approaches are insufficient. Conventional modelling requires extensive compound-specific data, whereas machine learning can identify patterns, capture complex interactions, and make predictions even in the absence of complete datasets. This project aims to develop a rapid, cost-effective machine learning model to predict the distribution and the fate of PFAS in ecosystems, accounting for key exposure pathways and environmental variables. The model will support regulatory agencies and scientific working groups by improving risk assessment and informing mitigation strategies.
Researcher(s)
Research team(s)
Project type(s)