Older urban landscapes present unique and complex stressors to urban streams and their habitats through the introduction of legacy and emerging toxic contaminants. Contaminant sources are often associated with various developed land uses such as older residential areas, active and former industrial sites, contaminated sites, and effluents from municipal wastewater treatment plant discharges. These landscapes have a history of legacy contaminant use such as polychlorinated biphenyls (PCBs) resulting in impacts to sediment and water in these complex environments. Despite the ban of PCBs in new commercial use in 1979, PCB contamination is still widespread in the environment, with many fish consumption advisories throughout the Chesapeake Bay region based on elevated PCBs. Several watersheds in the Baltimore region have mandated reductions in PCBs per total maximum daily loads in tidal waters of the watersheds in order to promote compliance with water quality standards. Some of these mandated reductions (for example, regulated watershed runoff) specified in the total maximum daily loads are the responsibility of the local jurisdictions as part of their phase 1 National Pollutant Discharge Elimination System municipal separate storm sewer system permit. In cooperation with the Baltimore City Department of Public Works and Maryland Department of the Environment, the U.S. Geological Survey and University of Maryland, Baltimore County conducted a study from 2018 to 2020 to refine the sources of PCBs from the City of Baltimore into Back River and to use the results to improve the conceptual site model of PCBs in the Back River watershed.

PCB concentrations in the water column of the nontidal streams in Back River watershed are relatively consistent throughout both tributaries, with greater concentrations detected in samples collected from Moores Run but greater loads estimated in samples collected from Herring Run. PCB concentrations measured in the bed sediments and analysis of the flux between sediment porewater (hereafter porewater) and surface water within the tributaries suggest that there are no stationary legacy sources within the stream channels.

The bulk of PCB mass entering the system from these nontidal tributaries appears to be introduced primarily during storm events. While only one storm event was sampled and concentrations were quantified only in Herring Run, solids captured during the storm were characterized by increases in PCB mass and overall suspended solids concentrations. Although the bioavailability of the PCB-associated sediment is unknown, this mechanism appears to warrant additional attention to better understand how concentrations vary under different storm conditions and temporally. The importance of contaminated stormwater in loading to Herring Run is further supported by the PCB concentrations in storm drain sediments collected near the tributary, which were present in higher concentrations and were characterized by different homolog signatures compared to that in bed sediments.

The observations in the tributaries differed from PCB concentrations and sediment characteristics downstream from the City of Baltimore boundary, in the upper tidal area of the main stem of Back River, particularly at the passive sampler locations BRT–1 and BRT–3. This depositional environment is characterized by higher organic content in sediments and higher concentrations of PCBs in porewater, which result in the possible flux of contaminants from sediment to the water column. This flux is generally opposite of that observed in the nontidal tributaries and the farthest upstream tidal site (BRT–2) and may be a result of the possible settling of sediment particles introduced via suspended solids in stormwater.

Despite an observed considerable reduction in overall PCB mass loading to and from the Back River Wastewater Treatment Plant (BRWWTP) (and similar reductions observed in biosolids) compared to the estimates previously reported from 2015, effluent from the BRWWTP continues to be a primary source of PCBs to Back River. The current study confirmed the likeliness of fat, oil, and grease deposits within the miles of sewer pipe as a source of PCBs to the BRWWTP influent. The differences between PCB concentrations in fat, oil, and grease deposits found in pipes (during replacement) compared to that of the BRWWTP suggest that legacy deposits may contain higher PCB concentrations and may act as a source of PCBs to passing sewage, eventually entering the BRWWTP. Variation in freely dissolved concentrations in the sewer system was apparent through the analysis of PCBs in the primary pump stations using passive sampling, with the largest contribution to the influent attributed to a single pump station and associated piping.

The contribution of PCBs to Herring Run and Moores Run via sanitary sewer overflows compared to the BRWWTP effluent is negligible, similar to reports from another large urban wastewater treatment plant. Therefore, decreased occurrence of sanitary sewer overflows is not expected to largely decrease PCB loads.

Results of this study suggest that targeted, sediment-capture best management practices in Back River watershed could be an effective way to reduce PCB mass loading assuming that deposited contaminated sediments are effectively isolated. Recent studies of some common urban best management practices such as bioretention have shown removal of PCBs within the stormwater control structures. In addition, appropriately timed street sweeping practices with appropriate collection equipment may be an effective way to reduce contaminants such as PCBs from road runoff sources. Reductions in concentrations and mass loading within the sewer system measured in this study compared to that estimated 5 years prior reflect the possible success of ongoing gray infrastructure management actions. Reductions may be attributable to enhanced nutrient reduction upgrades to the BRWWTP and extensive capital improvements and maintenance to the sewer system.

This study employed a combined sampling approach and a variety of sampling methods to include low-density polyethylene passive samplers, high-volume water samples, and grab samples of both water and sediment to characterize the PCB inputs to Herring Run, Moores Run, and Back River. Incorporating the passive samplers provided a time-weighted average of the freely dissolved concentration in the surface water, porewater, WWTP influent and effluent, and pump station influent over the deployment period with picogram per liter detection limits. A similar monitoring approach from this study could be implemented within other subwatersheds or municipal separate storm sewer system jurisdictions to assist in refining primary sources of PCBs in order to inform appropriate mitigation approaches.