Every Pound of Nitrogen

IThe argument

The Chesapeake Bay restoration is the largest water-quality experiment ever run in the United States, and its central result is now in. The deadline set in 2010 — every practice in place by 2025 — passed with nitrogen at 57 percent of goal, phosphorus at 90 percent, and sediment at 100 percent. On 2 December 2025 the deadline was moved to 2040. That is not a failure of effort or of money; roughly two billion dollars a year moves through the program. It is a failure of allocation, and the ledger that follows shows where.

Four facts organise everything below. First, since 2009 wastewater treatment plants delivered 64 percent of all nitrogen reductions and 61 percent of phosphorus reductions, at roughly forty dollars a pound; that work is finished, and Maryland's 66 upgraded plants alone now keep 11.7 million pounds of nitrogen a year out of the water — as much as the State's entire credited reduction since 2009. Second, the sector that is growing, runoff from developed land, is being treated with the most expensive practices in the catalogue: the median Maryland stormwater retrofit removes a pound of nitrogen for $1,558 a year, the median Maryland farm practice for $16 — a hundredfold gap measured by the same university on the same invoices, and confirmed by EPA's own program office at $2,259 against $24. Third, the watershed adds about 13,000 acres of pavement a year; Maryland's ten largest jurisdictions, spending more than $950 million over their permit terms, treat about 3,900 a year. Fourth, the remaining nitrogen gap is not everywhere: it is 33 million pounds, and on the last full accounting Pennsylvania alone was 29 million pounds short while Maryland's whole shortfall sat on the Eastern Shore, in nine agricultural counties.

The conclusion is uncomfortable for a region whose stormwater industry, our own included, has been built on the retrofit. Measured by the Bay, the next dollar belongs on a Pennsylvania or Eastern Shore farm, in a septic tank, or in the air. Measured by the neighbourhood, the stormwater retrofit still earns its keep — for flooding, stream stability and local rivers, and increasingly for rainfall that the models say will come 11 to 19 percent harder — but not for the Bay's nitrogen ledger. Programs that keep pretending otherwise will keep buying the least nitrogen per dollar in the watershed, and the deadline will move again. The scientists who wrote the program's own 2023 evaluation put the pace of improvement at a quarter of a percentage point of the Bay a year; at that rate the standards are three and a half centuries away.

IIThe water

The Bay holds about 18 trillion gallons and receives about 51 billion gallons of fresh water a day — 18.6 trillion a year, so the estuary turns over roughly its own volume annually. Half arrives down the Susquehanna; the Susquehanna, Potomac and James together carry more than 80 percent. The land draining to it covers 64,000 square miles in six states and the District, at a land-to-water ratio of roughly fourteen to one, the largest of any coastal watershed in the world. Whatever falls on that land, and whatever is on it, reaches the Bay with very little in between. The long-term mean inflow is 79,300 cubic feet a second; water year 2019 set the record at about 130,000, and the swing between a wet year and a dry one — 84 billion gallons a day in 2019 against 42 in 2023 — is larger than any change policy has produced.

Nobody publishes a single figure for how much stormwater the watershed's pavement generates, so we computed one and mark it as ours. The 2017/18 high-resolution land cover counts 1,927,940 acres of impervious surface, 4.75 percent of the land. One inch of rain on one acre is 27,154 gallons. At forty inches a year the region's roofs, roads and parking receive about 2.1 trillion gallons of rain — a ninth of the Bay's volume, 113 days of river inflow — and nearly all of it runs off; a city block generates more than five times the runoff of a forested block of the same size.

Two things stand out. The volumes engineering has kept out of local rivers are real but small against the system: the District's tunnels, at roughly $2.7 billion the largest stormwater project in the region, have captured 18.9 billion gallons of combined sewage over their operating life — about nine hours of river inflow. And the volumes still going in are concentrated in a handful of old combined-sewer cities: Richmond alone discharges around 1.8 billion gallons of mixed sewage and stormwater a year into the James, on a legislated schedule to 2035 and a $1.3 billion bill.

How much of the watershed's pavement is treated at all is not known as one number. Maryland's ten Phase I jurisdictions, with 173,838 impervious acres between them, had restored 46,782 acres across two permit terms by mid-2024 — 27 percent of their baseline. Prince George's County's public-private partnership has retrofitted about 5,000 acres since 2015. The District's retention-credit market captures about 40 million gallons a year. Those are the largest programs in the watershed, and together they treat a low single-digit share of its impervious surface.

IIIWhat is contaminated

The Bay's problem is not, in the main, toxic. It is arithmetic: too much nitrogen and phosphorus feed algae, the algae die and decompose, and decomposition strips oxygen from deep water each summer. In 2025 the watershed delivered an estimated 252.7 million pounds of nitrogen, 13.2 million pounds of phosphorus and 17.6 billion pounds of sediment to tidal waters. Agriculture is the largest source; developed land contributed about 16 percent of nitrogen, 15 percent of phosphorus and 10 percent of sediment in 2024; air deposition accounts for roughly a third of nitrogen once traced to the land it falls on.

The outcome measure the whole program is built around is the share of tidal water meeting its dissolved-oxygen, clarity and chlorophyll standards. It was about 26.5 percent in 1985–87 and 29.4 percent in 2021–23. In between it climbed to a record 42.2 percent in 2015–17 and fell back after the record rains of 2018–19. The program's own 2023 evaluation put the underlying rate of improvement at a quarter of a percentage point a year.

The dead zone has been kinder lately for reasons that have little to do with policy: dry springs. The 2023 hypoxic volume was the smallest on record; 2025 was near the median and ended early when Hurricane Erin mixed the water column; the 2026 forecast, issued in May, called for a volume 31 percent below average because January–April nitrogen loads were 39 percent below normal and river flow 32 percent below normal. A dry year is not a clean year, and it should not be scored as one.

Contamination in the narrower sense is nonetheless pervasive. In 2024, 80 percent of the Bay's tidal segments were partially or fully impaired by toxic contaminants — PCBs above all, with mercury and, increasingly, PFAS — down only from 84 percent in 2018. Maryland's 2024 Integrated Report lists 607 waterbody-pollutant combinations on its impaired-waters list, with 196 new temperature listings and 36 for PFOS; the report states plainly that all segments of the Bay are covered by nutrient TMDLs and that almost all remain impaired for nutrients. PCBs drive two-thirds of Virginia's 94 fish-consumption advisories. These are legacy loads that no buffer strip removes.

IVThe ledger, year by year

The Total Maximum Daily Load that EPA issued on 29 December 2010 is a pollution budget with a due date. Its original allocations — 185.9 million pounds of nitrogen, 12.5 million of phosphorus, 6.45 billion of sediment — were written on the Phase 5.3 model; after the watershed model was rebuilt, the operative 2025 planning targets became 219.5 million, 12.8 million and 18.6 billion pounds. The two sets are not comparable, and much public confusion comes from mixing them. Even within the current model the 2009 baseline is restated every year as inputs are revised (297.9 to 297.8 to 297.1 million pounds), which is why the annual series below should be read for direction, not for decimals.

Who did the cutting matters more than the totals. The Bay Program's own attribution: since 2009, wastewater upgrades produced 64 percent of nitrogen reductions and 61 percent of phosphorus reductions. The program's Scientific and Technical Advisory Committee put it more sharply for 2009–2021: wastewater 65 percent of nitrogen and 76 percent of phosphorus; cleaner air a further 25 percent of nitrogen; agriculture and urban land, which hold three-quarters of the remaining controllable load, together only 8 percent of nitrogen and 12 percent of phosphorus — about 3 million pounds of nitrogen a year from a decade of nonpoint work. Two sectors went the other way.

VMaryland, county by county

Maryland is the one jurisdiction in the watershed that publishes enough to build a county ledger, and it is the one that matters most to a Mid-Atlantic contractor. The State reports its nitrogen at basin scale: five basins, 45.8 million pounds of target, 46.6 million of load in 2024. Three basins are already under target. The whole of the shortfall is the Eastern Shore — nine counties, 17.8 million pounds against 15.6 — with small deficits in the Patuxent and the Susquehanna fringe.

Read together, the county tables say something the statewide numbers hide. The counties with the money and the permits — Montgomery, Prince George's, Baltimore, Anne Arundel, Howard — are the counties whose basins are already under target. The counties with the load — Kent, Queen Anne's, Caroline, Dorchester, Wicomico, Somerset, Worcester — have no MS4 obligation, no stormwater fee, and a farm economy in which one integrator's manure decides the phosphorus balance. Maryland's $239 million a year of stormwater spending and its $30 million cover-crop program are aimed at different halves of the State, and only one half is short.

VIWhat is actually helping

Three kinds of evidence answer this question, and they do not agree. The model credits practices at their designed efficiency. The state programs report what they paid for. The rivers report what arrived. Where all three agree — point sources — the answer is unambiguous. Where they diverge — farm practices and stream restoration — the honest answer is "less than credited, for reasons we partly understand."

Point sources: unambiguous

Enhanced nutrient removal at treatment plants is the one practice for which the model, the invoice and the river all agree. Maryland's 66 of 67 major plants run at an average effluent of under 2.85 mg/L of nitrogen against a 3.25 goal; Blue Plains, at 370 million gallons a day the largest advanced plant in the world, has kept more than 144 million pounds of nitrogen out of the Potomac since it began nitrogen removal, and the Potomac's grass beds came back behind it. Mattawoman Creek, where a 1991–95 plant upgrade cut nitrogen 60 percent, saw chlorophyll fall from 50–100 micrograms a litre and grasses return within one to four years. When Baltimore's two plants failed in 2021 the State's wastewater nitrogen jumped from 5.46 to 7.98 million pounds in a year; when enforcement restored them, it fell back. The signal is immediate because the pipe is one pipe.

Air: the reduction nobody paid for

USGS's accounting of inputs from 1985 to 2019 finds atmospheric nitrogen deposition fell by about 7 kilograms per hectare across the watershed — roughly 163 million kilograms in total — as the Clean Air Act cut power-plant and vehicle emissions. STAC credits cleaner air with a quarter of all nitrogen reductions. No Bay agency funded it.

Farm practices: large on paper

The model's biggest agricultural credit, Maryland's cover crops, is worth 3.4 million pounds of nitrogen a year at 491,000 acres. The field science supports the practice and undercuts the credit: early-planted rye cuts nitrate leaching 67 percent, late-planted wheat 27 percent, and in Maryland from 2008 to 2017, 68 percent of cover-crop fields were wheat and 35 percent were planted late. Late termination after 1 May adds 48 percent more uptake; most fields are terminated earlier. Satellite verification now covers 98 percent of Maryland's cover-crop fields, which is a real advance, but it verifies presence, not performance. At the watershed scale STAC found that a decade of nonpoint work produced about 3 million pounds a year of nitrogen reduction, and USGS finds that agricultural nitrogen and phosphorus surpluses have been rising since 2009 in the eastern and northern watershed, most consistently in Pennsylvania. Groundwater lag compounds the problem: median groundwater age on the Delmarva is 20 to 40 years, so a buffer planted in Kent County today will show in the Chester River around 2050.

Stream restoration: the contested credit

About 266 miles of stream have been "restored" in the watershed since 2010, with 84 more planned — more than 600 permits in Maryland and 300 in Virginia since 2014, and roughly $500 million by one 2020 estimate. It is the largest structural practice in every Maryland county's stormwater plan. STAC's 2024 review found moderate water-quality improvement at some monitored sites, one — Flatlick Branch — exceeding its credited reduction, but that "significant biological uplift is rare," that reach-scale projects do not address watershed stressors, and that removing mature trees to regrade a channel causes short-term harm. Maryland's own scientists doubt "that we can bring the ecology back." The practice earns its credit under a protocol; whether it earns it in the stream is, by the program's own account, unsettled.

Urban retrofits: working, when maintained

The Bay Program's tracking of 2,886 urban practices found inspection documented for 63 percent in Maryland and the District against 98 percent in Virginia; older Northern Virginia and King County studies found over half of practices failing within five years for want of maintenance. Every credit in this study assumes a working practice. A rain garden that has silted in removes nothing at any price.

Oysters: small, real, and rising

Harris Creek's 350 restored acres hold more than two billion oysters that filter the creek every ten days in summer and have removed about a million pounds of nitrogen over a decade — roughly 100,000 pounds a year, or a third of the entire Chesapeake Bay Trust Fund's projected annual credit. Eleven tributaries and 2,400 acres are now restored; Maryland's oyster population has tripled since 2005; Virginia's shellfish farm-gate value rose from $53 million to $81 million between 2018 and 2024.

Buying pounds: the price signal

Where the region has let a verified pound of nitrogen find its price, the price has been low. Maryland's Clean Water Commerce program, funded at $20 million a year from the flush tax, bought 192,143 pounds a year in its second round at an average of about $21 a pound against a $150 ceiling — 19 percent cheaper than the first round — with the best farm projects at $14 to $16 and shoreline or wetland projects at $179 to $200. Virginia's offset fund charges $5.08. PENNVEST's 2024 procurement drew bids from 46 cents to $460. Maryland's registry holds 1.36 million pounds of certified credits waiting for buyers.

Effectiveness, fourteen ways

"What works" is not one question. It is at least fourteen: what a practice removes as a share of what reaches it; what it removes per unit built; what a dollar buys; what an acre costs; who has actually delivered the reductions; how long each pathway takes to show in the water; what the practice does in the field rather than in the model; whether it is inspected and whether it survives; what the pipe discharges before and after; what the largest programs have banked; what the rivers say; how much of the plan is in the ground; why counties choose the practices they choose; and what each county pays. The fourteen exhibits below take them in that order, each on its own scale, each with its source.

VIIWhat works per dollar

The best cost data in the watershed come from the University of Maryland Center for Environmental Science, which in 2021 priced actual Maryland projects — the invoices, not the estimates — annualised them over their design lives at 5 percent, and divided by the pounds the Bay Program credits. The medians: a Maryland stormwater practice removes a pound of nitrogen for $1,558 a year and a pound of phosphorus for $9,639; a Maryland agricultural practice for $16 and $489. EPA's program office, from fifteen years of watershed-wide cost profiles, reaches the same order of magnitude: about $24 a pound for agriculture, $354 for stream and shoreline restoration, $2,259 for urban practices — "nearly 100 times that for agriculture," in EPA's words. Maryland's Bay Cabinet, reporting to its legislature, puts wastewater at about $40 a pound and as low as $10 with the operating incentive, septic at $130–330, stormwater at "$384 to more than $10,000." Three institutions, three methods, one answer.

Within stormwater the spread is itself instructive. The cheapest structural practice per pound is the sand filter at $384, followed by infiltration on sandy soils, urban tree planting and bioswales in the $600s. Bioretention — the rain garden that has become the signature of the field — costs $1,558 on soils that drain and $6,226 on the Piedmont clays, where it needs an underdrain and behaves like a filter. Permeable pavement removes nitrogen at $31,404 a pound. Maryland's counties confirm the range from the other direction: their projected cost per impervious acre restored runs from $9,523 in Baltimore City to $260,050 in Howard, with the State median at $45,816.

The program-scale figure is starker. STAC found that Maryland's municipal permittees planned to spend $1.195 billion to remove 85,000 pounds of nitrogen a year — $14,000 per annual pound of capital, for 0.2 percent of the reduction the Bay needs. The 2014 USDA analysis found that paying farmers for delivered pounds rather than installed practices would cut the cost of the agricultural reduction from $1.26 billion a year to $55 million. Neither finding has been controversial for a decade; both remain, in the main, unimplemented.

VIIIWhat works per man-hour

Labour is where the published record thins out, and we say so. Nobody in the watershed publishes hours per pound. Two things can be measured honestly. The first is routine maintenance labour, which the University of New Hampshire tracked for two to four years on a set of full-scale practices: from 6 staff-hours per impervious acre treated per year for porous asphalt to 28.5 for a sand filter. Credit each practice at Maryland's default removal for its class and a pound of nitrogen costs between half an hour and four hours of a crew's time a year to keep flowing.

The second measurement is cruder and more revealing. If every dollar of a practice's annual cost were a wage — none of it is, but the ratio says how much human effort a pound of nitrogen commands — then at a loaded $30 an hour for a landscape or construction labourer in the Washington–Baltimore market (Bureau of Labor Statistics means of $20.55 and $22.18, plus burden), the farm median of $16 a pound is about half an hour of work, the treatment plant's $40 an hour and a quarter, the stormwater median of $1,558 fifty-two hours, and permeable pavement a thousand. A retrofit contractor should read that not as an indictment but as a statement of where the value it delivers actually lies: in the hours themselves, in the flood it prevents and the stream it stabilises, and not in the pound of nitrogen it is nominally paid for.

IXThe treadmill

The developed-land load rose about 1.5 million pounds of nitrogen a year between 2009 and 2023 while every other major sector fell, and the reason is on the ground. Between 2013 and 2021 the watershed added about 13,226 acres of impervious surface a year — twenty square miles annually, more than a hundred thousand acres over eight years — nearly two-thirds of it driveways, parking and other flatwork, fastest in the suburbs of Baltimore, Harrisburg, Richmond and Washington. Some 4,400 acres of solar arrays appeared in 2021–22 alone, two-thirds in Virginia. Community tree canopy fell a net 28,000 acres in the same period despite 17,082 acres planted. From 1990 to 2007 impervious area under single-family homes grew 34 percent while population grew 18 percent: the region paves faster than it grows.

Regulation on new construction is the only lever that scales with the paving. Maryland's Stormwater Management Act of 2007, in force since May 2010, requires Environmental Site Design to the maximum extent practicable; the District's 2013 rule requires sites over 5,000 square feet to retain the first 1.2 inches of rain or buy the difference as credits; Virginia's Chesapeake Bay Preservation Act has held a 100-foot buffer in Tidewater since 1991. By Maryland's own loading rates, forest kept as forest avoids about 18 pounds of nitrogen an acre a year against pavement (20.39 against 2.31), permanently, with no maintenance crew. The partnership has protected 9.2 million acres, 23 percent of the watershed; the revised Agreement asks for 2 million more by 2040.

XThe big players

The restoration is often described as a partnership of seven jurisdictions and a federal office. In money and in work it is a much shorter list.

XIThe long push, 1983–2040

The Bay effort is old enough to have a history, and the history explains the present. Each decade has ratcheted the machinery — a voluntary pledge, then a numeric goal, then a federal budget with a due date, then a court-tested one — while the biology has responded slowly and the politics have rewarded the cheapest pounds first.

1. 1972Clean Water Act. Section 117, added later, creates the Chesapeake Bay Program Office within EPA.
2. 1975–83Congress funds a $27 million, five-year EPA study at Senator Mathias's urging; it names nutrient over-enrichment as the primary cause of decline.
3. 1983First Chesapeake Bay Agreement: a one-page pledge by Maryland, Virginia, Pennsylvania, the District, EPA and the Chesapeake Bay Commission.
4. 1984–88Maryland's Critical Area Act (1984); Virginia's Chesapeake Bay Preservation Act (1988). USGS begins the River Input Monitoring network (1985); the first attainment assessment finds 26.5 percent of the Bay meeting standards.
5. 1987Second Agreement sets the first number: a 40 percent cut in nitrogen and phosphorus by 2000.
6. 1990–99EPA's Phase I (1990) and Phase II (1999) municipal stormwater permit rules.
7. 2000Chesapeake 2000: 102 commitments; Delaware and New York join by memorandum, West Virginia in 2002. The 40 percent goal is missed.
8. 2004Maryland enacts the Bay Restoration Fund — the flush tax — to pay for enhanced nutrient removal at 67 major plants.
9. 2007Maryland's Stormwater Management Act requires Environmental Site Design to the maximum extent practicable; in full force May 2010.
10. 2009Executive Order 13508 puts the federal government formally in charge of leadership and two-year milestones. The TMDL baseline year.
11. 201029 December: EPA issues the Chesapeake Bay TMDL, the largest ever, with 2025 as the deadline and 2017 as the 60 percent midpoint. Phase I Watershed Implementation Plans.
12. 2011–16The American Farm Bureau and home builders sue; the TMDL is upheld in district court (2013), by the Third Circuit unanimously (6 July 2015), and the Supreme Court declines review (29 Feb 2016). Maryland mandates stormwater fees in its ten largest jurisdictions (2012), raises the flush tax to $5 a month, then repeals the fee mandate (2015) in favour of biennial financial-assurance plans. Prince George's signs the Corvias partnership (2015).
13. 2014Chesapeake Bay Watershed Agreement: ten goals and 31 outcomes, from oysters to public access, with the TMDL as its clean-water spine. Maryland's 20 percent impervious-restoration permits take effect.
14. 2015–17Wastewater meets its 2025 limits ten years early. Attainment reaches its record, 42.2 percent (2015–17). Grasses reach 104,843 acres.
15. 2018–19Midpoint assessment: nitrogen reaches only 40 percent of the 2017 target. Record rainfall in 2018–19 produces the highest river flows since gauging began in 1937; monitored nitrogen doubles; grasses fall 40 percent. Phase III WIPs are filed; Pennsylvania's is $324 million a year short. Maryland's Court of Appeals upholds the 20 percent permits against Carroll and Frederick (Aug 2019).
16. 2020–21Maryland, Virginia, Delaware, the District and CBF sue EPA over Pennsylvania (Sept 2020). Baltimore's Back River and Patapsco plants fail; Maryland's wastewater nitrogen rises 2.5 million pounds. The Conowingo WIP is written to offset 6 million pounds. Maryland's new permits cut the restoration requirement to about 11 percent.
17. 2022–23Pennsylvania's Clean Streams Fund puts $220 million of federal relief into farm conservation. The Pennsylvania suit settles (June 2023). STAC's evaluation concludes the nonpoint program is "insufficient" and the deep water is barely responding. The smallest dead zone on record. The District's Northeast Boundary Tunnel opens; Anacostia overflows fall 98 percent.
18. 2024The 2023 progress run shows 57 percent of the nitrogen goal. STAC's stream-restoration review finds biological uplift "rare." The Beyond 2025 committee reports; the Executive Council orders the Agreement rewritten. The best report card in 21 years.
19. 2025The deadline year. Federal grant freezes and staff losses at the program office in the spring; the appropriation survives. Maryland settles Conowingo for $340 million (October). The SHA permit is reissued at 15 percent (August). On 2 December the Executive Council adopts a revised Agreement: four goals, 21 outcomes, new load targets by 2030, all practices by 2040, a 2033 midpoint. Maryland and Virginia had argued for 2035. Governor Shapiro of Pennsylvania takes the chair.
20. 2026Congress appropriates a record $93 million (January). The Phase 7 model is accelerated to 1 January 2027. The final 2025 scorecard, 6 August: 57, 90 and 100 percent. Maryland's four largest county permits expire in November.
21. 2027–33Phase 7 model (2027); a TMDL reconsideration STAC expects in 2027; Virginia's 100 percent MS4 deadline (Oct 2028); the District's permit (Dec 2028); revised load targets and Maryland's flush-tax statute change (2030); the Potomac tunnel (2030); the Agreement's midpoint (2033).
22. 2035–40Richmond's combined-sewer deadline; the Agreement's habitat outcomes (2035); STAC's proposed climate reassessment (2035); all practices in place (2040) — 57 years after the first Agreement.

XIIWhat the future will do

The program has now modelled its own future in enough detail to be quoted. What the models say, in order of confidence: it will rain more and harder; the water will warm; the sea will rise; the population will grow; the loads will rise from climate alone; the dead zone will grow unless the loads fall faster than the climate pushes them up; and the cheapest reductions have already been made.

The policy calendar

The revised Agreement's numbers, as adopted or as drafted for comment, are smaller than the ones they replace. Tidal wetlands: 1,000 acres restored and 15,000 enhanced by 2035, against 85,000 and 150,000 in 2014. Forest buffers: 7,500 acres a year by 2035. Community trees: 35,000 acres by 2035. Oyster reef: 2,000 more acres by 2040. Land: 2 million more acres by 2040. Water quality: "continue implementing through 2030" and revise the targets by then, on the Phase 7 model due 1 January 2027. The program's own outcome memo of November 2024 projects, at 2009–2023 rates, nitrogen attainment around 2036 and phosphorus around 2034 — which is to say the 2040 deadline is the current trend, not a plan to beat it.

Money is the variable. The ACE Act authorises $92 million a year for the program office through FY2030 and Congress has so far appropriated above it; the 2027 request would cut the state revolving fund that built the treatment plants by 90 percent. Maryland's flush tax faces a statutory change on 1 July 2030 and its Trust Fund's competitive grants are being cut by a third. Pennsylvania's gap is $324 million a year; over the fifteen years to 2040 that is $4.9 billion the State has not yet found. Virginia's 100 percent MS4 requirement lands in October 2028; Maryland's four largest permits are renegotiated this autumn, with the SHA permit's 15 percent read by the counties as the template. No jurisdiction has yet published a cost to reach the 2040 targets. The Chesapeake Bay Foundation's 2014 estimate of the whole Blueprint was about $5 billion a year; extrapolated to 2040 that is $75 billion, against a claimed $22.7 billion a year of additional natural benefits when the Bay is restored.

XIIIWhat we would do

The evidence supports a program that would look very different from the one the region runs, and we would argue for it in this order.

1. Buy pounds, not practices. Every jurisdiction with a stormwater obligation should be allowed — and where its permit forces it, required — to meet a substantial share of its Bay nutrient obligation by purchasing verified reductions from the cheapest source in its basin, at prices Maryland's own purchases show are around $20 a pound. Maryland holds 1.36 million pounds of certified credits with too few buyers. Restricting trading is a decision to pay a hundred times more per pound.
2. Finish the point sources, then the septics. Princess Anne, Richmond's combined system on its 2035 schedule, the Potomac tunnel in 2030, and the conventional septic systems within reach of tidal water are the last large, permitted, verifiable pounds in the watershed at $40–330 each — and the one sector where the river confirms the model.
3. Pay farmers for measured outcomes, concentrated where the load is. Lancaster and York in Pennsylvania, Sussex in Delaware, and Maryland's nine Eastern Shore counties hold nearly all the remaining controllable nitrogen. Buffers, fencing, manure transport and cover crops cost $3 to $16 a pound. Verification is the whole problem — STAC is explicit that credited practices have not shown up in the streams as expected, and cover crops planted late as wheat deliver a third of their credit — so pay on monitored edge-of-field and small-watershed results, not on installed acres, and accept the twenty-year groundwater lag rather than pretend it away.
4. Hold the line on new pavement, and size for the rain that is coming. The one practice that costs the public nothing is not building the runoff. ESD standards, on-site retention rules like the District's, forest-conservation offsets and canopy targets should be treated as the primary stormwater program, and the retrofit as the remedy for what those rules failed to prevent. Adopt MDE's 11 percent larger design storm now rather than after Atlas 15 makes it unavoidable; Frederick and Charles are about to be built.
5. Re-justify the retrofit on its real merits. Stormwater retrofits in the built environment are worth doing — for flooding, for stream channels eating into back yards and sewer lines, for the Anacostia and the Jones Falls and the Elizabeth, for canopy and heat. Those are local benefits, and local governments should score, fund and defend them as such. Charging them to the Bay's nitrogen account at $1,558 a pound makes them look like the worst investment in the watershed when they are in fact a different investment. Within the retrofit budget, prefer sand filters, infiltration, swales and trees on soils that drain; be honest that bioretention on Piedmont clay is a filter that costs four times as much; stop building permeable pavement for nutrient credit; and hold stream restoration to an ecological outcome, not a protocol.
6. Fund maintenance as a first-class line. Every credited pound assumes a working practice. Maryland and the District document inspections on 63 percent of tracked practices. A program that builds and does not maintain has, on the historical failure rates, a half-life of about five years. Triennial inspection, a maintenance reserve of 8–13 percent of capital a year, and certified crews are cheaper than the rebuild — and are the part of this industry's work that is unambiguously worth its hours.
7. Stop scoring dry years. The Bay Program should report attainment and hypoxia flow-adjusted, so that a 31-percent-below-average dead zone in a 32-percent-below-average flow year is reported as what it is. Fifteen years of deadlines have taught the region to celebrate weather.

A note on our interest. This firm's principal works in the stormwater and landscape construction trade, and the retrofit is part of how that trade makes its living. We have written the numbers as the sources give them. The conclusion that the retrofit's value lies outside the Bay's nitrogen ledger is not a retreat from the work; it is the argument for funding it honestly.

XIVWhat we could not verify

The following were sought and either not found in a primary source, found only in secondary coverage, or found only in a form we could not read. None is used as a load-bearing figure.

• A jurisdiction-by-jurisdiction table for the 2025 progress run. The Bay Program's 6 August 2026 release gives watershed totals and, in a quotation, Maryland's 99/100/100; Pennsylvania's, Virginia's and the others' 2025 figures are not in text.
• Maryland's county-level 2025 nitrogen and phosphorus targets. They exist in the Phase III WIP's Appendix C, but pages C-3 to C-49 are page images and the State's server refuses downloads to this environment; the basin table is used instead. The 2012 Phase II county targets exist but are on a superseded model and are not comparable.
• Impervious acreage for the six Phase II counties; septic-system inventories and unit counts by county (the fund reports dollars); Howard's and Charles's current fee rates; the Trust Fund's dollars by county (the DNR report shows it only as a chart).
• The full attainment series for 1986–88 through 2011–13, which is published only in the interactive indicator; VIMS hypoxia values before 2019 on the current model vintage; MD DNR summer-average hypoxia for most years before 2020; SAV totals for 1985–2001 and 2003–04.
• The Bay Program's own press release for the 2021/22 land-cover totals (the 13,226 acres a year is reported by the Bay Journal and Maryland Matters from USGS's 2024-edition data); impervious acres by state.
• The 2026 UMCES report card (2025 data) and the 2025 SAV survey, neither published at the time of writing; the 2026 Executive Council meeting.
• Verbatim per-hectare maintenance hours for five of the seven Houle practices (the range 14.8–70.4 is verbatim; the per-acre values are from the authors' EPA presentation).
• EPA Chesapeake Bay Program appropriations for FY2009–2017 and FY2022–23 in text form; Maryland's state Bay spending before FY2022 as an annual table.
• The six practices and per-pound costs in the Chesapeake Bay Commission's 2004 Cost-Effective Strategies (a scanned PDF); cost-per-pound figures for green roofs, precision feeding or Clean Air Act rules; the "$30–40 billion" cost-to-finish figure sometimes attributed to the Commission, for which no source was found.
• Labour hours per acre for cover-crop planting, riparian planting with containerised stock, or street-sweeping productivity per curb mile; firm-level revenues or project totals for the region's restoration contractors; poultry-integrator market shares.
• Station-specific NOAA 2022 sea-level values for Baltimore, Annapolis and Norfolk (regional Northeast values are used); Tian et al. (2022) hypoxia figures; the precise content of the 2030 amendment to Maryland's Bay Restoration Fund statute, which we describe as a "change" rather than a sunset.
• The Bay Program's "2024 BMP verification findings" as a document; Anne Arundel's pay-for-success stream outcomes; Virginia's statewide MS4 reductions and the SLAF cumulative total; Pennsylvania's PRP percentages from DEP's own pages (its site blocks fetching).

XVSources

Method. Figures are quoted as published; dollar-years are stated where sources give them and are not inflated to a common year. Where we computed a figure it is marked as ours in the exhibit note. Each exhibit's data is available as a CSV of the name given in its note; the full set is filed with the research package. Second edition, September 2026: adds the year-by-year series, the Maryland county ledger, the methods and players sections, and the projections, and corrects the first edition's per-acre maintenance hours, its street-sweeping label, and the attribution of the District's captured-overflow figure. Higgins Strategy is registered and disclosed; one principal is in every meeting; we never take both sides of a matter.