Distinguishing natural and anthropogenic stressors at a groundwater-surface water interface using a weight-of-evidence framework.

Abstract

This study applied a tiered weight‑of‑evidence (WoE) framework to evaluate whether contaminants of potential concern (COPCs)-chloride, ammonia, total dissolved solids, and elevated pH-transported through groundwater-surface water interactions contribute to benthic impairment at a calcium chloride production site. The assessment focused on five test stations (A to E) near suspected discharge zones and upstream reference sites within a drowned river‑mouth wetland, allowing comparison of exposure and biological responses across a hydrologic gradient. Four lines of evidence (LOEs) were integrated to distinguish natural and anthropogenic stressors influencing benthic macroinvertebrate communities: (1) Physicochemical characterization of COPCs; (2) benthic community assessments using indices of biotic integrity based on the Rapid Bioassessment Protocols and the Great Lakes Coastal Wetland Monitoring Program, along with reciprocal recolonization studies; (3) in situ toxicity testing; and (4) laboratory microcosm tests simulating contaminant upwelling under reasonable worst-case (RWC) conditions. Surface water pH remained within acceptable limits, whereas chloride and ammonia exceeded the Michigan EGLE regulatory thresholds at stations D and E. Ammonia concentrations exceeded the acute criterion at station E and the chronic criterion at station D. Both stations also experienced nighttime hypoxia (<1 mg L-1 DO). In situ toxicity tests implicated low dissolved oxygen (DO) and elevated ammonia as primary stressors, and microcosm experiments confirmed ammonia-related toxicity under RWC conditions. Benthic bioassessment indices and recolonization studies indicated moderate to severe impairment at stations D and E, while upstream and less-affected stations exhibited communities consistent with reference conditions. A fourth-phase investigation further confirmed DO depletion as a dominant stressor at station E, which showed the highest contaminant levels and the most impaired benthic communities. Overall, these findings highlight the value of a site‑specific WoE framework for distinguishing natural variability from legacy contamination in complex wetland environments and show how iterative refinement improves stressor identification and reduces misinterpretation.