PFAS Fate and Transport Modeling

As part of Michigan State University’s Institute of Water Research, our PFAS program advances the science of contaminant monitoring, modeling, and prediction at watershed, regional, and statewide scales. Our work integrates hydrology, geochemistry, machine learning, and advanced geospatial analytics to support Michigan, the Great Lakes region, and national partners in addressing PFAS contamination with scientific rigor and practical tools.

Priority Areas

Synthesis of PFAS Knowledge Gaps: Comprehensive reviews of PFAS fate processes across surface water, soils, groundwater, and biological systems to identify critical research needs in integrated large-scale modeling.
Surface Water–Groundwater Connectivity: Coupled SWAT–MODFLOW–WASP modeling frameworks to assess PFAS fluxes via runoff, lateral flow, groundwater discharge, and sediment dynamics, with demonstrated applications in PFOS-impacted watersheds.
Integrated PFAS Fate & Transport Modeling: Development of a distributed, watershed-scale PFAS model that simulates movement through surface soil, vadose zone, rivers, and lakes, enabling source attribution and mitigation planning.
Statewide PFAS Prediction in Drinking Water: Machine learning and conditional autoregressive models that estimate PFAS occurrence and concentrations across Michigan, achieving high accuracy while reducing data requirements.
Next-Generation Deep Learning for Monitoring: First-of-its-kind application of Graph Neural Networks (GNNs) for PFAS prediction in unsampled public supply wells, capturing complex geospatial, lithologic, and network relationships.

Present Issues

Persistent and Complex PFAS Pathways

PFAS move through interconnected hydrologic systems via runoff, erosion, leaching, groundwater flow, and sediment transport. Their adsorption at air–water and water–solid interfaces and exceptionally slow degradation make them long-term contamination drivers.

Legacy Emissions and Diffuse Sources

Wastewater treatment plants, industrial facilities, biosolid-amended fields, landfills, airports, and historical fire-training activities introduce long-chain PFAS such as PFOS and PFOA into soils, aquifers, and streams. Identifying the relative contributions of these sources requires integrated models and improved monitoring.

Data Limitations for Decision-Making

Many watersheds lack sufficient PFAS sampling density. This limits the ability to track contaminant plumes, quantify historical loads, and anticipate future risk, especially for communities that rely heavily on groundwater.

Modeling Gaps

Existing tools often simulate isolated parts of the system; few capture soil–vadose–groundwater–surface water interactions simultaneously. Substantial opportunities remain to improve sorption processes, precursor transformation, and transport under variable climate and hydrogeologic conditions.

Our Vision

We aim to protect water resources and public health by providing scientific insights and advanced analytics to agencies, communities, tribal nations, utilities, and researchers. Our vision includes:

Integrated Environmental Modeling: Unified frameworks linking surface water, unsaturated soils, groundwater, and stream networks to capture PFAS interactions across the full hydrologic system.
Predictive Intelligence: AI, graph neural networks, and scalable statistical models to identify PFAS hotspots, forecast concentrations, and guide monitoring investments.
Human–Water System Integration: Science that connects hydrologic processes to social and policy outcomes, enabling adaptive management under uncertainty.
Decision Support & Transparency: Open dashboards, reproducible analytics, and communication tools to help practitioners interpret PFAS patterns and evaluate remediation or sampling strategies.

Team Leadership

Team Lead

Dr. A. Pouyan Nejadhashemi — Director, Institute of Water Research (IWR)

Michigan State University Foundation Professor and Director of IWR, Dr. Nejadhashemi leads a cross-disciplinary PFAS research program spanning watershed modeling, groundwater dynamics, risk analytics, and environmental decision support. His work has produced:

The first watershed-scale integrated PFAS model.
Statewide predictive frameworks for PFAS occurrence and concentration in public drinking water.
Advanced models for predicting PFAS in unsampled wells with quantified uncertainty.
Foundational syntheses of PFAS fate, transport, and modeling knowledge gaps.

Tagline: PFAS Modeling, Hydrologic Systems, Machine Learning, Environmental Decision Support.

Selected Publications

Raschke, A. L.; Nejadhashemi, A. P.; Rafiei, V. Overview of Modeling, Applications, and Knowledge Gaps for Integrated Large-Scale PFAS Modeling. J. Environ. Eng. 2022, 148 (9), 05022005. DOI: 10.1061/(ASCE)EE.1943-7870.0002033
Raschke, A. L.; Nejadhashemi, A. P.; Rafiei, V.; Fernández, N.; Shabani, A.; Li, S. Opportunities and Challenges of Integrated Large-Scale PFAS Modeling: A Case Study for PFAS Modeling at a Watershed Scale. J. Environ. Eng. 2022, 148 (9), 05022005. DOI: 10.1061/(ASCE)EE.1943-7870.0002034

Rafiei, V.; Nejadhashemi, A. P. Watershed Scale PFAS Fate and Transport Model for Source Identification and Management Implications. Water Res. 2023, 240, 120073. DOI: 10.1016/j.watres.2023.120073

Fernández, N.; Nejadhashemi, A. P.; Loveall, C. M. Large-Scale Assessment of PFAS Compounds in Drinking Water Sources Using Machine Learning. Water Res. 2023, 243, 120307. DOI: 10.1016/j.watres.2023.120307

Rafiei, V.; Nejadhashemi, A. P. Graph Neural Networks for PFAS Prediction in Water Supply Wells. Eng. Appl. Artif. Intell. 2026. 32, 112746. DOI: 10.1016/j.engappai.2025.112746.