A reusable, code-based way to pinpoint which tributaries keep adding PFAS to a river

What is it about?

Finding where PFAS "forever chemicals" enter a river usually means sampling above and below each segment and describing the pattern in words. That makes results hard to reproduce or compare across trips. This study introduces a machine-readable "reach bracketing" framework: the chemistry (what was measured at each site, including non-detects) is kept in one table, and the river layout (which sites are upstream, tributary, and downstream of each other) is kept in a separate, reusable table. A short script then reads the river-layout table and computes PFAS changes along each reach, with uncertainty quantification, so anyone can rerun or reuse the analysis. We applied it to two winter sampling campaigns in the Trinity River headwater tributaries near Fort Worth, Texas (December 2025 and February 2026), covering 15 sites and 40 PFAS analyzed by EPA Method 1633. One small tributary, Farmers Branch (site TW91, next to Naval Air Station Joint Reserve Base Fort Worth), stood out dramatically: its total PFAS reached about 15,667 ng/L, roughly 20 times higher than the next site, and it carried a chemical signature consistent with aqueous film-forming foam. That elevated signal reappeared seven weeks later in the second campaign, so it appears to be a persistent input rather than a one-time event. We also provide a free browser-based tool so any monitoring program can run the same framework on its own data.

Featured Image

Photo by Dan Roizer on Unsplash

Photo by Dan Roizer on Unsplash

Why is it important?

Most PFAS river studies report concentrations at individual sites but describe upstream-to-downstream relationships only in narrative form, making those comparisons hard to audit, reproduce, or reuse. This work encodes the river network itself as a reusable, machine-readable object, so reach-scale contrasts become explicit and transferable across campaigns, analysts, and watersheds, consistent with FAIR data principles. It pairs that structure with censoring-aware uncertainty estimates suited to the high non-detect rates typical of EPA Method 1633 data. For agencies and utilities, it offers a low-cost reconnaissance tier that flags priority reaches, like the persistent Farmers Branch input near NAS JRB Fort Worth, before committing to more expensive, discharge-based source investigations. The same approach can be applied to other low-level contaminants, such as pesticides, pharmaceuticals, and trace metals.

Perspectives