The Pipe Bergen County Can't See Into

In Tenafly, the pipe that carries rainwater off the streets and the pipe that carries sewage from homes are supposed to be separate. That is how the town's system is built, and how most of Bergen County's is built too. But a hard rain does not respect that separation. Water seeps through cracked sanitary pipes and leaky manholes and shows up, uninvited, as sanitary flow during storms, even in a system engineered to keep it out. Utility engineers have a name for this: rainfall-derived infiltration and inflow, or RDI/I. It is Tenafly's own core infrastructure problem, and it sits at the center of a much larger regional question. Bergen County's sewers cannot see what is happening inside them in real time, while twenty minutes away, New York City's can.

That gap, between pipes that were sized decades ago and the cheap, open-source sensors now watching similar pipes elsewhere, is the story of what has and has not changed in how the region manages wastewater and stormwater since Bergen's sewers were laid.

Two Pipes, One Problem

Most of Bergen County runs what engineers call a separate sewer system, meaning stormwater and sanitary sewage travel through different pipes to different endpoints, at least on paper. The Bergen County Utilities Authority (BCUA) treats sanitary flow for 47 Bergen municipalities, processing more than 83 million gallons of wastewater a day at its two treatment plants, while the state's Municipal Separate Storm Sewer System (MS4) permit program, run by the New Jersey Department of Environmental Protection, governs what towns must do with the stormwater side.

Older urban cores nearby do not have that separation. Newark, Paterson, Hoboken, Jersey City, and New York City's five boroughs still run combined sewers, where stormwater and sewage share a single pipe on the way to a treatment plant. The Passaic Valley Sewerage Commission (PVSC), formed in 1902, treats wastewater for 1.5 million people across 48 municipalities in Bergen, Essex, Hudson, Passaic, and Union counties.

It moves that flow through a 21-mile Main Interceptor Sewer running from Paterson to its Newark Bay treatment plant. North Bergen, a Bergen County municipality, is one of PVSC's eight combined-sewer towns, along with Newark, Paterson, Jersey City, Bayonne, Kearny, Harrison, and East Newark. That makes Bergen County's exposure to combined sewer overflows direct, not a downstream concern borrowed from someone else's infrastructure. New York City DEP, for comparison, operates 14 wastewater resource recovery facilities and a combined sewer network covering roughly 60 percent of the city.

A combined sewer overflow, or CSO, happens when heavy rain pushes more flow into a combined pipe than the treatment plant can handle, forcing the excess, untreated sewage mixed with stormwater, to discharge directly into a river or bay. During moderate storms, those eight communities send that untreated mix into the Passaic River, the Hackensack River, Newark Bay, the Kill Van Kull, and New York City's waterways. Bayonne alone operates 28 separate CSO outfalls, each equipped with solids and floatables controls, with 16 discharging to Newark Bay, 9 to the Kill Van Kull, and 3 to Upper New York Bay. That single municipality's outfall count gives some sense of how many individual discharge points the regional system as a whole is trying to manage.

RDI/I is a quieter version of the same underlying problem, one that does not produce a visible overflow into a river. Instead, it shows up as unexplained sanitary flow spikes during storms, straining treatment capacity in towns that believe their systems are already separated from stormwater. As Bergen towns densify, older single-family lots replaced by larger footprints with more paved and roofed surface, both RDI/I and peak flow get worse without a single new pipe being installed anywhere.

More Rules, Same Old Pipes

New Jersey has spent the past several years rewriting the rules that govern how new development handles stormwater, in direct response to the fact that rainfall itself has changed. In 2023, NJDEP's Inland Flood Protection Rule required new development to design for future storm conditions rather than historical ones, and every New Jersey municipality had to update its stormwater ordinance by July 2024 to comply. Around the same period, roughly 100 New Jersey municipalities that had been classified as Tier B under the state's Municipal Separate Storm Sewer System permit program were reclassified as Tier A under the 2018 MS4 permit update, a shift that brought stricter federal Clean Water Act obligations to towns with some of the oldest infrastructure in the state and no new funding attached to help them meet it.

That effort has since been folded into something larger. On January 20, 2026, NJDEP adopted the Resilient Environments and Landscapes (REAL) rules, a sweeping amendment package covering the state's Coastal Zone Management rules, Freshwater Wetlands Protection Act rules, Flood Hazard Area Control Act rules, and Stormwater Management rules. The REAL rules replace the old design flood elevation standard with a Climate-Adjusted Flood Elevation, built on climate modeling that projects conditions through the year 2100, and require new or substantially improved structures to be built four feet above FEMA's base flood elevation. On the stormwater side specifically, the rules now require water quality treatment for redevelopment projects with motor vehicle areas even when no new impervious surface is added. They also redefine reconstruction of a quarter acre or more of impervious surface as major development, triggering full stormwater design and volumetric reduction requirements.

The rules carry a legacy provision: projects with a complete application submitted to NJDEP within 180 days of the rule's effective date, meaning by July 20, 2026, can still be reviewed under the prior standards. That window closes in ten days. After that, every new Bergen County project, from a Tenafly teardown to a larger commercial redevelopment, gets designed against the new climate-adjusted standard rather than the old one. The rules have not gone unchallenged. Senate Continuing Resolution 106 and Assembly Continuing Resolution 59 both allege that NJDEP lacks the statutory authority to adopt them, an unresolved legislative fight running in parallel with the rules' rollout.

The REAL rules also lean toward nature-based solutions over pipe-and-concrete fixes where possible, prioritizing green infrastructure, wetlands preservation, and permeable surfaces as tools for managing runoff before it reaches a pipe at all. That preference reflects a six-year rulemaking process that started with the original NJ PACT executive order in 2020 and moved through the 2023 Inland Flood Protection Rule before landing on the current package.

None of this rewrites the pipes already in the ground. The REAL rules govern how new development manages stormwater going forward. They do nothing for the thousands of miles of existing sanitary pipe already leaking rainwater into the system, or for towns trying to figure out where in their own network RDI/I is worst without digging up every block to find out. New York City has at least started building that kind of visibility into its own network. NYC DEP's data-driven catch basin inspection program uses prioritization software, rather than a blanket cleaning schedule that treats every catch basin the same regardless of condition, and has cut resolution times by 30 percent as a result. Bergen County has no equivalent program for its own storm side of the system.

The Sensor Nobody Has Ordered

That is where a different kind of infrastructure enters the picture, one that does not replace a single pipe but tells engineers, in real time, what the pipes they already have are doing.

FloodNet is a sensor network built by researchers at NYU Tandon and CUNY, including Brooklyn College's Science and Resilience Institute at Jamaica Bay and the CUNY Advanced Science Research Center. The project runs in partnership with the NYC Mayor's Office of Climate & Environmental Justice, NYC DEP, and the city's Office of Technology and Innovation. Each solar-powered sensor uses an ultrasonic range-finder to measure water depth to within an inch, transmitting a reading every minute over LoRaWAN or cellular networks. Deployment costs run under $300 per sensor. Since a 2020 pilot, the network has grown to more than 350 sensors across New York City's five boroughs as of April 2026. That expansion has been funded by $7.2 million in city money, with a goal of reaching 500 sensors by the end of 2026. Every sensor design is open source on GitHub, alongside a free public dashboard anyone can check during a storm, showing which sensors are actively recording a flood event and how deep the water has gotten at each location. Separately, NYC DEP has installed 241 sewer monitors citywide measuring water levels and, in some locations, flow, in what the department has described as a step toward smart sewers, though this monitoring runs alongside FloodNet rather than replacing it.

FloodNet's own researchers describe the underlying problem as one most residents outside the affected neighborhoods do not know is happening at all. NYU Tandon's Andrea Silverman, the project's principal investigator, has pointed to neighborhoods around Jamaica Bay that flood as often as seven times a month, a frequency chronic enough that it functions less as an occasional disaster and more as a routine feature of daily life for the people who live there. That is the specific kind of recurring, low-grade flooding the sensor network was built to make visible, both to the residents deciding whether to take a different route to a subway stop and to the city agencies deciding where to send limited water-rescue resources during a storm.

Even NYC's system is still, for now, a monitoring layer rather than an actively controlled one. The more advanced stage, real-time control, or RTC, uses actuated valves and gates to physically redirect flow away from overloaded parts of the system as a storm happens, adjusting the network the way a traffic engineer might retime signals during rush hour rather than waiting to rebuild the road. A handful of other American cities have already reached that stage. South Bend, Indiana's CSOnet combines 120 sensors with 12 actuators and has sharply cut the city's combined sewer overflow discharge.

Evansville, Indiana's AI-driven real-time decision support system has cut combined sewer overflows by more than 100 million gallons a year, at roughly 5 percent of the cost of building conventional storage tanks to hold the same volume. In Detroit, the Great Lakes Water Authority, working with University of Michigan researchers, has projected an additional 100 million gallons of effective sewer capacity from dynamic control alone, worth an estimated $500 million in avoided construction costs. Beyond RTC itself sits a further stage nationally: AI-based predictive control, where a system forecasts where and when capacity will be exceeded before a storm arrives, rather than reacting to sensor readings as the water rises. Evansville's system already leans in that direction, but the fully predictive version of the approach has not been tried anywhere in this region at all.

None of that exists in Bergen County or the broader northern New Jersey suburbs right now. There is no RTC deployment and no FloodNet-style sensor network anywhere in the county, a striking absence given how close it sits to a system that has already proven the concept. The absence is not simply a matter of the technology being unproven or expensive. Drexel University engineer Franco Montalto has been running FloodNet sensors in Camden, New Jersey, since 2019 under a National Science Foundation grant, using the data to calibrate hydrologic and hydraulic flood models.

That deployment sits inside New Jersey, funded and active for years, which turns "why not Bergen County" from a hypothetical into a directly answerable question. The technology has, in fact, traveled further than that. A University of Georgia-led team is deploying FloodNet-style sensors in San Juan, Puerto Rico, under a National Science Foundation grant, building toward what the project's lead researcher calls impact-based forecasting, translating a rainfall forecast into a specific prediction of which streets will flood and when. Similar open-hardware sensor networks are now spreading into Brazil. The hardware is open source, inexpensive, and has already been proven to work in a New Jersey city. It has simply not yet been ordered by anyone in Bergen County.

Who Pays, and Who Decides

If a lack of urgency were the only obstacle, it might be a simple matter of a phone call to BCUA. The harder version of this problem is that fixing the CSO side of the system, the side that produces a visible, permitted discharge into a named river, is not straightforward even where everyone agrees it should happen, because agreeing on who pays for it is its own fight.

PVSC's Long Term Control Plan, developed with program manager TYLin, evaluated two funding tiers for reducing combined sewer overflow events across its district: roughly $1.2 billion to cut overflow events to about 20 per year, or roughly $1.9 billion to cut them to about 4 per year. PVSC tested those alternatives using a calibrated, system-wide InfoWorks Integrated Catchment Model, a hydraulic-hydrologic modeling platform that simulates how a given fix would actually perform before a dollar gets spent on construction. That same modeling produced an uncomfortable finding for a story that would otherwise end with the CSOs eliminated and the rivers clean. In the Upper Passaic River, Lower Passaic River, and Upper Hackensack River, PVSC's modeling found that water quality standards for pathogens cannot be met even with complete elimination of combined sewer overflows, because background sources, including other stormwater runoff, upstream discharges, wildlife, and dry-weather sources, are already high enough on their own. Fixing every CSO in the district would not, by itself, make those rivers meet the standard.

A separate piece of gray infrastructure illustrates how tightly the whole regional system is linked together. A proposed secondary treatment bypass expansion at PVSC's Newark Bay plant would let the system reliably treat up to 720 million gallons per day of wet-weather flow, but that expansion depends on upgrading the pumping and conveyance capacity of the Hudson County Force Main, a single piece of infrastructure gating the performance of the entire regional system.

PVSC has run a formal public engagement process around its Long Term Control Plan since 2016, when it formed a Supplemental CSO Team made up of members of the affected public to act as a liaison between the utility and residents in the eight combined-sewer communities. That team, operating under the branded outreach name Clean Waterways, Healthy Neighborhoods, meets quarterly and has also funded municipal green infrastructure feasibility studies alongside its work on the control plan itself. The engagement structure exists. Whether it has actually shifted how the cost of fixing the CSOs gets divided among the communities that host the infrastructure is a separate question, one the 2025 permit comment period brought back into the open.

NJDEP issued renewed draft CSO permits in 2025 for PVSC and its eight combined-sewer municipalities, with a public comment period running into late October 2025. The comments filed during that window show how unresolved the cost question remains.

The South Ward Environmental Alliance, a Newark environmental justice group, calculated that Newark's individual municipal cost under one funding scenario would run to $92.1 million, with each of the eight CSO communities owing roughly $27.3 million toward a shared $219 million allocation, on top of their respective share of a separate $45 million cost tied to the proposed second bypass. The group's comments state plainly that all eight CSO communities are environmental justice communities, and separately call on NJDEP to formally require use of the EPA's Environmental Justice Screening tool in future permitting decisions rather than leaving that identification informal. The letter also draws a direct comparison that undercuts any argument that this problem is simply too complicated to solve: it points out that, unlike PVSC's plan for Newark, the Camden County Municipal Utilities Authority did build an equity-focused approach into its own CSO planning, weighing the burden on the host community rather than treating it as a fixed cost of doing business. The group's comments argued that Newark and the other seven communities are being asked to bear a construction burden, along with the odor and truck traffic that come with it, disproportionate to their share of the regional benefit, while the towns and cities that send flow into the system without hosting any of the treatment infrastructure pay comparatively less. This is the same procurement-equity dynamic that has shown up elsewhere in this series: the places that host infrastructure are not always the places that most directly benefit from it, and the funding formula rarely accounts for that gap on its own.

What's Still Unresolved

Bergen County's own path forward is, at this point, less a funding fight than an absence of a proposal to fight about. No agency has put forward a plan for what a Bergen County RTC or sensor pilot would cost or where it would go, which means the county is not stuck in a funding debate so much as a step behind having one. Whether the REAL rules' new stormwater crediting standards make it easier or harder for a small town like Tenafly to address RDI/I on its own system is also an open question, since the rules were written with new development in mind rather than retrofits to existing sanitary pipe. And the legislative challenge to NJDEP's authority to adopt the REAL rules in the first place remains in play, meaning the regulatory foundation for how the region designs its way out of these problems is not entirely settled either.

What is clear is that the tools to see inside Bergen County's sewers already exist, are already funded and running twenty miles away in Camden, and cost less than a home security camera to install. What has not yet happened is anyone deciding to buy one.