The distinction between combined and separated sewer systems is central to understanding why Philadelphia faces this challenge. In a separated system, sanitary sewage travels through one set of pipes to a treatment plant while stormwater runoff is conveyed through a separate network that discharges directly into local waterways. In a combined system, both flows share the same conduits. When rainfall is modest, the full volume of combined flow reaches the treatment facility. When rainfall is intense or prolonged, however, the volume can exceed the capacity of both the pipes and the treatment plant, and the excess — a mixture of stormwater, untreated sewage, and whatever debris and pollutants have accumulated on streets and rooftops — is released through designed overflow points into the nearest receiving waters. This design choice, common in cities built in the nineteenth and early twentieth centuries, was intended as a safety valve to protect streets and basements from flooding, but it means that heavy storms routinely deliver diluted sewage directly to rivers and creeks.

The scale of the problem in Philadelphia is substantial. More than 8 billion gallons of combined sewer and stormwater discharged annually represents a chronic, recurring input of organic matter, nutrients, pathogens, and other contaminants into the city’s waterways. Because these discharges are tied to rainfall patterns, any trend toward more frequent or more intense storms — a pattern widely projected for the northeastern United States — translates directly into more frequent overflow events. That connection between precipitation trends and overflow frequency is what makes the long-term perspective of the Drexel analysis particularly valuable: a decade of monitoring captures the natural range of wet and dry years, allowing the researchers to separate the seasonal rhythms of dissolved oxygen from the disruptions imposed by overflows.

Dissolved oxygen deserves particular attention as a monitoring target because of its dual role as both a direct ecological constraint and an integrative signal of watershed condition. Aquatic organisms vary in their oxygen requirements, but cold-water fish species and many sensitive macroinvertebrates need consistently high concentrations to survive, grow, and reproduce. The Environmental Protection Agency’s minimum thresholds are set to protect these uses, and sustained excursions below them can compress habitat into oxygenated refuges, alter food web structure, and in extreme cases produce fish kills or hypoxic zones devoid of most aerobic life. The biochemical oxygen demand exerted by organic-rich sewage is the key mechanism here: microorganisms decomposing the organic load consume oxygen, and the depletion can persist for days after the initial discharge has passed downstream. This lagged, cumulative character is why the researchers describe dissolved oxygen as recording a video rather than a snapshot — it integrates stresses over time in a way that instantaneous measurements of, say, bacterial counts or turbidity cannot.

The seasonal pattern identified in Tacony-Frankford Creek is especially instructive. Warm summer water holds less oxygen at saturation than cold winter water, so baseline dissolved oxygen levels are naturally lower in July and August. When an overflow event adds an oxygen demand during this already-stressed period, the water has less buffering capacity, and levels can fall below protective thresholds more easily and recover more slowly. The finding that 40 percent of July observations in that watershed fell below the EPA minimum, with low levels persisting through September, suggests that summer overflows may impose ecological stress that compounds across the warm season rather than washing away with each individual storm. By contrast, watersheds like Pennypack Creek and Darby-Cobbs Creek, which rarely dipped below the threshold at any time of year, appear to possess some combination of hydrological dilution capacity, land cover characteristics, or flow dynamics that buffer them against comparable depletion.

The contrast between the Delaware River and the smaller creek watersheds illustrates a principle that applies broadly to receiving-water ecology: volume and flow rate govern dilution, and dilution governs impact. A large river with substantial discharge can disperse an overflow plume quickly, spreading the organic load across a much greater volume of water and maintaining oxygen exchange with the atmosphere across a large surface area. A small urban creek receiving the same plume experiences a far higher concentration of organic matter per unit volume, and its limited reaeration capacity — the rate at which oxygen from the atmosphere dissolves back into the water — may be insufficient to replenish what the decomposing sewage consumes. Reaeration itself depends on turbulence, channel morphology, and flow velocity, all of which are shaped by the engineered condition of urban streams. Channelized, hardened sections of creek typically have different reaeration characteristics than naturalized reaches, which is one reason the researchers point to surrounding infrastructure and geography as important determinants of vulnerability.

Land cover within a watershed exerts its own influence on both the frequency and the consequences of overflows. Impervious surfaces — roads, parking lots, rooftops — prevent rainfall from infiltrating into soil, so a higher fraction of each storm arrives rapidly in the sewer system as surface runoff. Watersheds dominated by impervious cover therefore generate flashier hydrographs, with sharp peaks in flow that more readily trigger overflows. Conversely, tree canopy, vegetated ground, and permeable surfaces intercept rainfall and slow its arrival at the sewer network. This is the logic behind the mitigation strategies the researchers note: retention basins, permeable pavement, rain gardens, cisterns, and rain barrels all work by capturing stormwater where it falls and either releasing it slowly or allowing it to soak into the ground, thereby reducing the volume and peak rate of flow entering the combined system during a storm. Green stormwater infrastructure of this kind has been deployed extensively in Philadelphia as part of the city’s broader effort to reduce overflow frequency, and the Drexel findings suggest that its effectiveness may differ from watershed to watershed depending on each system’s underlying characteristics.

The watershed-by-watershed framing carries practical implications for how municipalities prioritize investments. Cities facing regulatory obligations to control overflows typically must choose among expensive options: expanding treatment capacity, building deep storage tunnels to hold combined flow until it can be treated, deploying distributed green infrastructure, or some combination. Because the ecological damage of an overflow depends on the receiving water’s capacity to absorb it, two overflows of identical volume can have very different consequences. A discharge into a large, fast-flowing river during a cool season may cause negligible oxygen depletion, while the same discharge into a small, warm, slow creek in late summer could push dissolved oxygen below survival thresholds for days. Data-informed prioritization — directing resources toward the most vulnerable receiving waters and the most ecologically sensitive seasons — can therefore achieve greater ecological benefit per dollar than a uniform approach. The Drexel model, built on more than a decade of USGS monitoring paired with overflow records, offers a template for the kind of analysis that supports such prioritization.

The composition of the overflows themselves remains a significant unknown. Combined sewer overflow discharges carry not only organic matter but also nutrients such as nitrogen and phosphorus, fecal microorganisms, suspended solids, metals washed from surfaces, hydrocarbons, and whatever contaminants have accumulated in the sewer between storms. Each of these constituents has different dynamics: nutrients can stimulate algal growth whose subsequent decomposition drives the oxygen demand the researchers measured; pathogens pose direct public health risks to recreational users; solids can smother benthic habitat. Marques notes that more monitoring and testing of overflow water could identify which areas face the greatest ecological risk, and this compositional detail matters because mitigation strategies differ in what they remove. Green infrastructure that captures the first flush of a storm may remove a disproportionate share of pollutants accumulated during dry weather, while storage-based approaches may be more effective at capturing large-volume events. Understanding what is actually in the discharge, and in what concentrations across different storm sizes, would allow cities to match interventions to the specific threats their receiving waters face.

The lagged nature of dissolved oxygen response also has implications for monitoring design. Programs that sample only immediately after storms may miss the deepest oxygen depletion, which can occur days later as microbial decomposition proceeds. Continuous sensors, which the USGS monitoring network deploys at many stations, capture these delayed minima and the full trajectory of recovery. The recovery period itself is ecologically meaningful: repeated cycles of depletion and recovery, occurring with each significant storm through a summer, may prevent sensitive species from reestablishing even if each individual event falls short of causing a fish kill. Chronic, sublethal stress of this kind can reshape community composition over years, favoring tolerant organisms and eroding biodiversity in ways that are difficult to reverse.

Looking ahead, the interplay between the built environment and overflow resiliency that the Drexel group plans to explore touches on one of the most consequential questions for aging cities: whether investments in greening and stormwater capture can measurably improve the ecological condition of receiving waters, and under what conditions. As precipitation patterns continue to shift and urban development alters impervious cover, the baseline against which overflow impacts are measured will itself move. Long-term monitoring programs like the one underpinning this study provide the record needed to detect such changes and to evaluate whether mitigation is working. For the many American cities whose combined sewers date to the 1800s, the Philadelphia analysis offers a concrete demonstration that the path to healthier urban waterways runs through an understanding of each watershed’s particular vulnerabilities — its size, its land cover, its flow regime, and the infrastructure that connects its streets to its streams.

Subject of Research: …

Article Title: How do combined sewer overflows affect water quality in Philadelphia?

Article References: How do combined sewer overflows affect water quality in Philadelphia?. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: combined sewer overflows, water quality, dissolved oxygen, Philadelphia, Drexel University, urban watersheds, stormwater management, Journal of Environmental Management, aquatic ecosystems, aging sewer infrastructure, combined, sewer