Temporal trends of environmental contaminants in a Durham, NC, creek using non-targeted analysis

Abstract

Third Fork Creek is an urban tributary in Durham, North Carolina, that drains to Jordan Lake, a regional drinking water supply. Because the last comprehensive chemical survey of this creek was conducted in 2014 and demonstrated poor water quality, we used a combined targeted and non‑targeted analysis (NTA) approach to establish an updated baseline of organic contamination and to prioritize analytes for future monitoring. We sampled three sites monthly from July to December 2022. Water samples were filtered, extracted by solid‑phase extraction, and analyzed by ultra high-performance liquid chromatography (UHPLC)–Orbitrap high-resolution mass spectrometry (HRMS). In total, 1,346 features were detected (1,043 in positive mode; 303 in negative mode). Spectral library matching returned 43 candidate identifications (29 positive; 14 negative), and mass‑list screening provided additional tentative identifications. Principal component analysis showed that temporal (monthly) variation explained more of the compositional differences than spatial location across the three sites. Detected compound classes included pharmaceuticals, per- and polyfluoroalkyl substances (PFAS), plasticizers, biocides, and personal care product residues. Compounds confirmed with analytical standards included atrazine, deisopropyl‑atrazine, metolachlor, imidacloprid, sucralose, dibutyl phthalate, caffeine, 6:2 fluorotelomer sulfonate, perfluorooctanesulfonic acid (PFOS), and 4‑methylbenzotriazole. Sucralose, a wastewater tracer, ranged from 1.2 to 53.5 nanograms per liter (ng/L) and was highest at the South Alston site in August and October. Suspect screening highlighted several additional priority contaminants for follow‑up, including 8‑hydroxyquinoline, sildenafil, and oxymetazoline. These results demonstrate that seasonal NTA campaigns can (1) provide a broad, up‑to‑date contaminant baseline in urban surface waters; (2) reveal temporal pulses and hotspots that routine targeted networks may miss; and (3) generate prioritized target lists to make subsequent routine monitoring and remediation efforts more efficient and cost‑effective.