InterPore2025

Reconstructing historical PFAS concentrations in groundwater using a reduced-order modeling framework

20 May 2025, 09:50

15m

Oral Presentation (MS18) Innovative Methods for Characterization, Monitoring, and Remediation of Contaminated Soils and Aquifers MS18

Speaker

Russ Detwiler (UC Irvine)

Description

As the incidence of PFAS-contaminated drinking water increases around the world, understanding of sources and pathways of PFAS in complex urban watersheds continues to develop. In the Orange County groundwater basin in Southern California, recent sampling of over 500 monitoring and production wells has revealed concentrations of a group of PFAS compounds that exceed federal health advisory levels. As a result, dozens of production wells have been removed from service while treatment systems are being built. The groundwater basin is actively managed and recharged with surface water from the local Santa Ana River (SAR), imported surface water and purified recycled water. PFAS has recently been measured in the SAR at concentrations consistent with wells surrounding the recharge facilities suggesting that historic concentrations of PFAS in the SAR may be a primary input of PFAS in the groundwater basin.
Here we use an extensive set of PFOS, PFOA, PFNA, PFBS, PFHpA, and PFHxS groundwater concentration measurements from across the basin to reconstruct the water quality history of the SAR for each compound. We represent SAR-to-well transport using a reduced-order modeling approach. We develop transfer functions that describe transport along one-dimensional streamlines between each well and the likely source of recharge. We invert for the source history using truncated singular value decomposition. Dispersion coefficients and travel times to individual wells are determined by a hierarchy of models and data that include a calibrated MODFLOW model of the basin and a set of chloride breakthrough observations. Uncertainties in historical SAR concentrations are quantified by a Monte Carlo analysis that includes the uncertainty inherent in the required model parameters. We show that our reconstructed histories are in close agreement with recent measurements of PFAS in surface water from the SAR. We also combine the modeling framework with historical estimates of historical source concentrations to hindcast groundwater concentrations at production wells over the past 30 years. Estimated concentrations generally fall within the uncertainty bounds predicted by the Monte Carlo analysis. Our results may be helpful for estimating historical exposures and guiding investments in treatment facilities.