Mitigation & Nutrient
Market Quarterly

Sales

Commercial mitigation banks made 75 sales between July and September: 34 of stream credits, 35 of non-tidal wetland credits and 6 of tidal wetland credits. That matches the rest of the year (72 in Q1 2026, 73 in Q2 2026) and is above the average of the previous four quarters (68).

Stream credits did best. There were 34 stream sales, the most in any quarter since Q2 2024. Together they came to 17,709 LF, 26% more than the four-quarter average. The average stream sale was 521 LF. Wetland sales were steady at 45.4 acres, about 1.3 acres per sale.

Advance sales are reported in batches. Three advance sales are recorded for the quarter, against 20 to 24 a quarter in 2025. Advance sales reach SWaN in batches, often many months after the sale. Of the 26 recorded since late July, 20 arrived together on 3 August, covering sales from April 2025 to March 2026. Observed

Sales took place in 19 watersheds. Middle Potomac-Catoctin had the most stream sales (8, totalling 1,835 LF). The largest amounts of stream credits were sold in Appomattox (3,105 LF), Middle James-Buffalo (2,167 LF) and Pamunkey (2,151 LF).

Credits for Sale

On 30 September, banks had 136,289 LF of stream credits and 416.3 acres of wetland credits for sale. Section 01.2 shows how much is open to any buyer, and where it is.

Between 3 August and 1 October, banks released 12,569 LF of stream credits and 43.9 acres of wetland credits for sale. Over the same period, 17,594 LF and 55.1 acres left the market, mostly through sales. As a result, the amount for sale fell slightly: by 3.6% for stream credits and 2.6% for wetland credits.

Most of the new stream credits came from one watershed. Middle James-Buffalo released 7,714 LF, 61% of the total. That is where Bridgeport Mitigation Bank opened on 8 July; it sold 2,167 LF before the quarter ended. Most of the new wetland credits (34.1 of 43.9 acres) came from Middle James-Willis.

Banks

On 30 September, Virginia had 214 stream and wetland bank records: 118 approved, 51 awaiting approval, 41 sold out, 3 suspended and 1 withdrawn. Twenty of them are single-client banks. Seventeen records were added to SWaN during the quarter, but only one was a new proposal. The others were older proposals that SWaN loaded from RIBITS, the US Army Corps of Engineers' bank register. Observed

Sellers

The quarter's 75 sales came from 26 banks, registered to 17 accounts. In the previous quarter there were 25 accounts, and a year earlier there were 24. The five largest accounts made 64% of sales, the highest share in any quarter since the start of 2024. In the previous ten quarters it ranged from 41% to 59%. The 17 accounts were also the fewest in that period, against 20 to 28 before. Observed

A concentration index. The Herfindahl–Hirschman Index (HHI) measures how concentrated sales are. It squares each seller's percentage share and adds the results, so it runs from close to 0 (many small sellers) to 10,000 (one seller). By seller account, the quarter's HHI was 1,161, the highest since the start of 2024; the previous ten quarters ranged from 519 to 922. That is as if sales were split evenly among 8.6 sellers. By bank, the HHI was 539, a typical level for the period. The rise therefore came from accounts selling from several banks, not from any single bank growing. Derived

Measured by the amount sold rather than the number of sales, concentration was higher still: 1,732 for stream credits and 1,928 for non-tidal wetland credits. US competition authorities treat markets with an HHI above 1,800 as highly concentrated. That benchmark is given for context only. Buyers face local markets that a statewide figure does not capture (see below). One quarter's 75 sales also make the index volatile, so the trend over several quarters matters more than a single reading.

Owners and delegates. Many bank owners do not sell their credits themselves. They delegate the work to another organisation that acts on their behalf. 56% of the quarter's sales came from banks with such a delegate. Counted by delegate, the picture changes. TNT Environmental, Inc. acts for 19 banks, including 13 registered to the "David Jordan" account, and handled 27% of the quarter's sales. Clearwater Ventures LLC handled 21%. Concentration is higher on this view: an HHI of 1,556, against 1,161 by owner. Observed

Owners and managers together. An organisation can influence a bank if it owns it or acts on its owner's behalf, and RIBITS names both. Counting both, TNT Environmental, Inc. reaches the most of the quarter's sales. It manages 13 approved banks for others, including all 10 owned by the "David Jordan" account, and those banks made 27% of the quarter's sales. Clearwater Ventures LLC owns 8 banks and manages 3 more, reaching 21% of sales. Fallings Springs LLC owns 11 banks, the most of any owner, and reached 11% of sales. Observed

Following these links, TNT Environmental and the seven owners it acts for form one group. It reaches 13.5% of approved banks but 27% of the quarter's sales. Clearwater Ventures and the owners it acts for reach 11% of banks and 21% of sales. Treated as blocks, the market's HHI on sales is 1,573, against 1,161 by owner alone. Derived

Where influence is concentrated. Under the service-area rules each watershed has its own local market, and influence varies from one to the next. In 12 of the 24 watersheds with enough sales to measure, one organisation owns or manages the banks behind at least half of local sales over the last four quarters. TNT Environmental leads in the two northern Potomac watersheds, Middle Potomac-Anacostia-Occoquan and Middle Potomac-Catoctin, reaching 44% and 47% of local sales. Virginia Wetland Consulting leads in the upper and middle James watersheds, and RES in the south-east, from Lower James to Meherrin. In Lynnhaven-Poquoson, banks owned by Thomas Tye account for 89% of local sales. Derived

These measures show where influence could sit, not that it is used. Owners linked only through a common manager have not necessarily acted together, and SWaN records that a manager may act for an owner, not the terms of that arrangement.

Local markets are much more concentrated. Credits can only be used within a bank's service area, so the sellers a buyer can turn to are those whose banks' service areas cover its watershed. Counted that way, every watershed with sellers active in the last four quarters is highly concentrated, with an HHI above 1,800: all 28 for stream credits and all 30 for non-tidal wetland credits. A typical watershed is served by just two active sellers; in six watersheds for stream credits and twelve for wetland credits, only one. Derived

SWaN records the account that registered each bank. That is usually the bank's owner, but not always. Where the account is in a person's name, the company behind it is not recorded. The account registered as "David Jordan" holds 14 bank records (10 approved, 4 awaiting approval), more than any other account. The next largest are Fallings Springs LLC (13) and Clearwater Ventures LLC (11). This account also registered Copper Hill Mitigation Bank, which was approved on 17 September and made its first sale before the end of the month.

Resource Environmental Solutions made four stream sales, totalling 3,922 LF. That was the largest average stream sale of any seller.

Buyers

91% of the quarter's sales name a buyer, and 51% can be linked to a known organisation. Most identifiable buyers were building transport, utilities or housing. Virginia Passenger Rail Authority made five purchases, more than any other buyer. Dominion Energy made three. Other buyers included Stafford, Chesterfield and Spotsylvania Counties, the City of Chesapeake, VDOT, and the housebuilders Toll Mid-Atlantic, Lennar and Stanley Martin. Toll's four purchases were recorded under three different spellings. Iron Mountain Data Centers made two purchases. Observed

The rail purchases match the permit record. DEQ issued a permit for the Third Track at Potomac Creek on 27 August (Section 01.5).

Buyers are far less concentrated than sellers. By the same index, the quarter's purchases had an HHI of 298 by number of purchases, as if they were spread evenly among 34 buyers; sellers were like 8.6. The five largest buyers made 28% of named purchases, against 64% for the five largest sellers. Measured by amount bought, buyer concentration was higher, because a few large purchases dominate: 799 for stream and 1,652 for non-tidal wetland credits. That is still below the sellers' 1,732 and 1,928. Derived

Within each watershed, buyers are concentrated too. Over the last four quarters, 11 of the 12 watersheds with at least three stream purchases, and 12 of the 14 for non-tidal wetland, had a buyer HHI above 1,800. With so few purchases in each watershed, this mainly shows how thin local markets are: in most watersheds, a handful of buyers deal with one or two sellers. Derived

Supply

On 30 September, banks had credits for sale in three products: stream, non-tidal wetland and tidal wetland.

Supply is concentrated in a few watersheds. Three hold 47% of the stream credits for sale: Middle James-Buffalo (24,181 LF), Pamunkey (22,904 LF) and Middle Potomac-Catoctin (17,395 LF). Non-tidal wetland supply is more concentrated still. Albemarle alone holds 221.6 acres, 53% of the state's total, and the next two watersheds bring the share to 71%. Almost all tidal credits for sale are in Hampton Roads, and about two-thirds of those belong to VDOT's single-client Goose Creek bank. Credits can only be used within each bank's service area, so where supply sits matters as much as how much there is. Observed

In-lieu fee programmes add a large amount of supply through advance credits, which they can sell before their restoration is done. Three programmes have advance credits available: the Virginia Stream and Wetland Mitigation Program (formerly VARTF), DEQ's Virginia Wetland and Stream Replacement Fund, and the Coastal Virginia Conservancy Trust. Together they have 293,544 LF of stream credits and 430.24 acres of non-tidal wetland credits available, with smaller amounts of tidal, subaqueous and oyster reef credits. For stream credits that is more than twice what banks hold; for non-tidal wetland it is slightly more. Advance credits are the fallback when no bank credits are available, so they matter most in the watersheds that are short of bank credits. But SWaN records advance credits as released to each programme as a whole, not to particular watersheds, and its records of advance sales do not say which programme made them. So this edition cannot show which watersheds the advance credits can serve, or how much each programme has left of each product. Derived

Demand

Demand comes from DEQ's official permit record, which SWaN has read directly since 30 September. The Editor's Note explains how far these figures can be relied on: they can be wrong in either direction. At the end of the quarter they stood as follows.

The last column matters most for the market. It shows the credits that permit holders still have to buy. The first two columns also include permits that are finished and fully paid for.

About a quarter of the requirement is still owed for both products: 25% for stream credits and 23% for non-tidal wetland credits. What is owed is concentrated in a few watersheds. Four hold 60% of the stream credits still owed: Lower Potomac (12,441 LF), Appomattox (9,468 LF), Nottoway (9,379 LF) and Rapidan-Upper Rappahannock (8,893 LF). For non-tidal wetland, Lynnhaven-Poquoson alone holds 56% (82.6 acres). In three watersheds for each product, more credits have been bought than are required. Observed

Where Supply and Demand Meet

Across the state, there are more credits for sale than are known to be owed. Banks have 416.3 acres of wetland credits for sale (about 265 acres excluding single-client banks), against 146.5 acres owed. They have about twice as many stream credits for sale as are owed. But two cautions apply. The figures for what is owed are not yet complete and could be wrong in either direction. And credits can only be used in the area each bank serves, so a statewide surplus does not mean every buyer can find credits nearby.

The trading rules decide what "nearby" means. The usual order is that a permittee buys from a bank whose primary service area covers the impact; if no such bank has enough credits, from a bank whose secondary service area covers it; and only then from an in-lieu fee programme's advance credits. So the market a buyer faces is the set of banks whose service areas reach its watershed, not the banks located in it. This section compares what is owed in each watershed with the credits held by those banks. A watershed counts as served by a bank when at least half of it lies within the bank's service area (Notes). Derived

Streams. Of the 17 watersheds where stream credits are owed, 12 are fully served by banks' primary service areas and one more by secondary service areas. Four are short, and together they hold 36% of the stream credits owed. No bank's service area covers Lower Potomac (12,441 LF owed), Conococheague-Opequon (2,978 LF) or Lynnhaven-Poquoson (127 LF). Rapidan-Upper Rappahannock is served by banks holding 5,176 LF, against 8,893 LF owed. Watersheds that look short when only local banks are counted, such as Appomattox and Nottoway, are well served by banks elsewhere: the banks serving Appomattox hold 21,603 LF, against 9,468 LF owed. Derived

Wetlands. Of the 24 watersheds where non-tidal wetland credits are owed, 13 are fully served by primary service areas and one more by secondary areas. Ten are short, and together they hold 67% of the acres owed. Most of that is in Lynnhaven-Poquoson: 82.6 acres owed, against 0.74 acres held by the banks that serve it. Meanwhile, 221.6 acres for sale, 53% of the state's total, are in the Albemarle watershed, where buyers have already bought more than they need. These wetland figures use the credits recorded for each bank, which do not match the snapshot totals exactly (Notes). Derived

Do Advance Credits Bridge the Gaps?

Where the banks serving a watershed hold fewer credits than are owed, the fallback is advance credits from in-lieu fee programmes. Four watersheds are short of stream credits, with a combined gap of 19,263 LF, and ten are short of non-tidal wetland credits, with a combined gap of 97.9 acres. Statewide, the programmes have 293,544 LF of stream and 430.24 acres of non-tidal wetland advance credits available: 15.2 times the stream gap and 4.4 times the wetland gap. Derived

Whether those advance credits can be used in each of these watersheds cannot be confirmed from SWaN's records, because advance credits are recorded for each programme as a whole. The sales record gives a strong indication. Since 2021, advance credits have been sold in all four stream gap watersheds and in 8 of the 10 wetland ones. In Lower Potomac, which no bank's service area covers, 23,960 LF of advance credits were sold, against 1,454 LF from banks: advance credits are in effect its market. In Conococheague-Opequon, all 7,697 LF came from advance sales. In Lynnhaven-Poquoson, one buyer bought 7.14 acres of advance credits in February 2026. Observed

The fallback has a published price. The programmes' fee schedules set an advance price for each watershed. For non-tidal wetland credits in the gap watersheds, it ranges from $70,000 to $300,000 per acre; in Lynnhaven-Poquoson it is $125,000 per acre. Banks' prices are not recorded, so whether the fallback costs more than a bank credit cannot be shown. Observed

In short: in total, advance credits are more than enough to cover the gaps, and they have been bought in nearly every watershed where banks fall short. What cannot yet be shown is whether each programme still has advance credits available for each of those watersheds. Derived

How Supply and Demand Are Moving

SWaN's daily record of credits for sale began on 3 August 2026, and its record of credits owed on 30 September. So this edition cannot show how those totals changed over earlier quarters; from the next edition, MNM will report them every quarter. What can be traced back is what moves them: banks opening and selling out, permits creating new requirements, and purchases meeting them.

Supply. Five commercial banks were approved this quarter, the most since July–September 2025 (nine), after one a quarter in the three quarters between. No commercial bank has sold out since late 2024. Even so, credits for sale ended the quarter lower than they started, because sales outpaced new releases: down 3.6% for stream credits and 2.6% for non-tidal wetland credits between 3 August and 1 October. Non-tidal wetland supply swung within that period, falling about 5% in late August and then rising above its starting level after releases at the end of the month. Observed

Demand. DEQ issued 83 VWP permits this quarter, the most in any quarter since the start of 2024; the previous ten quarters ranged from 37 to 74. Purchases held steady: 17,709 LF of stream credits and 45.4 acres of non-tidal wetland credits were bought from banks, close to recent levels. Observed

Taken together, new permits are adding to demand faster than at any time in the period, while new bank supply is only starting to pick up and the amount for sale is edging down. Whether that tightens the market will show in the next few quarters. Implied

None of the quarter's 75 commercial stream and wetland sales has a recorded price. The only prices on record are three fees for wetland credits from an in-lieu fee programme.

These are fees set by the programme, not prices agreed between buyer and seller. They show what the fallback option costs in each watershed, not what banks charge. They range from $60,000 to $100,000 per acre. In Q0, the only bank prices in the 2021–2025 period were two wetland sales at $65,000 and $95,000 per acre, both to public buyers.

The lack of prices is normal for this market, not a feature of this quarter. Banks sell privately, and RIBITS does not ask for a price. Prices become public mainly when a public body buys, because Virginia's procurement system, eVA, publishes the purchase. This quarter's public buyers included Virginia Passenger Rail Authority, VDOT and several counties. Their eVA records may show what credits cost, but those prices are not yet in SWaN. Directional

The Federal Proposal of 9 September 2026

The EPA and the US Army proposed new options for deciding which waters federal law protects, known as "waters of the United States" (91 FR 57284). The options add to a proposal they made in November 2025. Observed

Under one option, federal protection would cover only streams that have water all year round, allowing for one dry spell of up to 30 days and for drought. Under another, a wetland would be protected only where it shares year-round surface water with a protected water. Comments close on 9 October 2026. Nothing has been decided.

Why the Effect in Virginia Is Limited

Virginia's own permit programme, the VWP programme, has covered wetlands without needing a federal permit since 2000. That includes isolated wetlands that federal law may not cover. DEQ says federal changes to this definition do not replace its own authority. State law also requires developers to make up for wetland losses, so that there is no overall loss of wetlands. They can do this by buying bank credits or paying into a state fund.

So a narrower federal rule would move some projects from joint federal and state control to state control alone. Those projects would still have to make up for the damage. Implied We cannot tell from current data how much of today's demand comes from waters a narrower rule would exclude.

Demand on the Way

DEQ's record shows 83 VWP permits issued during the quarter. That compares with 37 in the previous quarter and between 59 and 74 in each of the four quarters before. Twenty-eight of them share the same date, 2 August, which suggests they were recorded in a batch. So the quarterly total is more reliable than the monthly figures. Directional Another 99 permits are at the application stage.

The permits issued are for transport, energy, data centres and building. Transport projects include VDOT roads, the Third Track at Potomac Creek, and security roads for the Metropolitan Washington Airports Authority. Energy projects include Southeast Virginia Energy Storage, Wildcats Solar, two substations and Reusens Hydroelectric. The two data centres are Wellington Glen and CXE01 Mullin. The rest are for housing and commercial development.

The applications include the redevelopment of a terminal concourse at Washington Dulles and the Claytor hydroelectric plant.

Each permit that requires credits is a demand signal. It tells the market how many credits of each type will be needed, and where, before they are bought. It becomes a purchase once the project goes ahead.

What to Watch

Sales

Nutrient banks made 410 phosphorus sales, totalling 551.8 lbs. The average sale was 1.35 lbs. That is 7% fewer sales than the average for the previous four quarters (443), and 12% less phosphorus (628 lbs). Sales have stayed between 410 and 450 a quarter for six quarters in a row.

A nutrient sale usually includes three credits: phosphorus, nitrogen and sediment. This report counts each sale once, by its phosphorus credit. This quarter, every phosphorus sale also included nitrogen, and 96% included sediment.

Northern Virginia's two Potomac watersheds had 35% of sales but only 11% of the phosphorus sold. Buyers there make many small purchases. Buyers in Hampton Roads and Middle James-Willis made fewer, larger purchases.

Credits for Sale

Phosphorus credits for sale fell from 6,128.9 lbs on 3 August to 5,960.1 lbs on 1 October. Over that time, banks released 483.8 lbs and 652.5 lbs left the market. More than half the new credits (279.1 lbs) were released in Hampton Roads. Smaller amounts came from Rapidan-Upper Rappahannock (74.0 lbs) and Middle Roanoke (45.8 lbs).

At this quarter's rate of sales, the credits for sale would last about eleven quarters. But that compares supply with past sales, not with future need. This market has no record of future need to compare supply against (Section 02.5). Derived

Banks

The register lists 410 nutrient banks: 250 approved, 137 sold out, 21 awaiting approval and 2 suspended. That is 34 more sold out than at Q0. The register does not record when a bank was approved or sold out, so we cannot tell how many did so this quarter. The 250 approved banks are spread across 37 watersheds.

Sellers

Resource Environmental Solutions (RES) owns CBAY-VA, LLC, as Q0 found from company records. Together, RES-linked banks made 98 of the quarter's 410 phosphorus sales (23.9%). Over the last four quarters their share was 23.1%. Banks linked to RES and NERD Farms together made 30.0% of sales.

36 of the quarter's sales came from 13 banks that SWaN cannot yet link to an owner, either because the owner or sponsor has not been identified or because the public record has discrepancies that are being resolved. These banks are not one seller, so this report counts each separately; counted as one, the quarter's HHI would be 1,100 rather than 1,032. SWaN records a delegation to Eco-Cap LLC for 8 of the 13, which made 30 of the 36 sales. In all, 80 banks made phosphorus sales this quarter, against 87 in the previous quarter.

Owners and delegates. Delegation is far more common in the nutrient market. 86% of the quarter's phosphorus sales came from banks whose owners have appointed another organisation to act on their behalf, and the largest seller changes when sales are counted that way. By owner, RES-linked banks made the most sales (23.9%). By delegate, Eco-Cap LLC handled the most (27%), ahead of RES (25%). Eco-Cap is linked to 154 nutrient banks. It acts for 145 under accepted delegations, including 13 of the 45 banks that cannot yet be linked to an owner. Eight more are registered to Eco-Cap itself, and one further delegation was awaiting acceptance at the end of the quarter. Many banks name both Eco-Cap and RES as delegates. Concentration is much higher on this view: an HHI of 1,715 by delegate, against 1,032 by owner. Observed

Owners and managers together. Counting both ownership and management, Eco-Cap LLC can influence the most nutrient banks. It owns 5 approved banks and manages 97 for others: 41% of the 250 approved. Those banks made 28% of the quarter's phosphorus sales. RES owns 22 approved banks and manages 36 for others: 23% of approved banks and 26% of sales. Many banks name both as managers. Observed

Following these links, 81 organisations form a single network: owners linked to Eco-Cap or RES, and the two managers linked to each other through the banks they share. The network reaches 60% of approved nutrient banks and 77% of the quarter's phosphorus sales. Treated as one block, the market's HHI on sales would be 6,084. That is an upper bound: owners linked only through a common manager have not necessarily acted together. Derived

Where influence is concentrated. Each watershed's local market includes its neighbours. Eco-Cap leads in 19 of the 45 local markets with enough sales to measure, across the James, York, Roanoke and Chowan watersheds; RES leads in 8 and Naturion in 7. In 26 of the 45, one organisation owns or manages the banks behind at least half of local phosphorus sales over the last four quarters: for example, Eco-Cap in Lower James (56%) and Appomattox (53%). Derived

These measures show where influence could sit, not that it is used. SWaN records that a manager may act for an owner, not the terms of that arrangement: whether a manager sets prices or chooses buyers is not known.

Counted by owner, phosphorus sales were less concentrated than stream and wetland sales this quarter: 1,032 by number of sales, as if they were split evenly among 9.7 sellers. But the index has roughly doubled since early 2024, when it was 551. Derived

Construction sites may use phosphorus credits from their own watershed or an adjacent one, so the local market is wider than for stream and wetland credits. Counted that way, a typical watershed has 14 active sellers within reach. Concentration is high in 26 of the 45 watersheds where phosphorus was sold, moderate (an HHI of 1,000 to 1,799) in 15, and low in 4. Derived

Buyers

Only 41% of this quarter's phosphorus sales name a buyer. That compares with 78% last quarter and 99% in early 2025, and it is the lowest share on record. So this quarter it is harder than ever to see who is buying. Directional

Among named buyers, VDOT (19 purchases) and Dominion Energy (18, plus 2 recorded as "Dominion VA Power") bought most often. Chesterfield County bought the most phosphorus: 64.9 lbs across 10 purchases. Amazon Data Services made two purchases.

Phosphorus buyers are the least concentrated group in either market. Before 2026 the buyer HHI was below 200 in every quarter, as if purchases were spread among more than 50 buyers. It was 348 this quarter by number of purchases. Only 41% of this quarter's sales name a buyer, so the index covers fewer than half of them. Derived

Buyers are measured within each watershed, where the purchase is made. Of the 37 watersheds with at least three phosphorus purchases, 21 had a buyer HHI above 1,800, measured on purchases that name a buyer. So phosphorus buyers typically choose among many sellers within reach, unlike stream and wetland buyers, who often have only one or two. Derived

Supply

On 1 October, nutrient banks had 5,960 lbs of phosphorus credits for sale, in 37 watersheds. Supply is less concentrated than for stream and wetland credits. The three largest holdings, in Upper Roanoke (1,009 lbs), Hampton Roads (835 lbs) and Upper New (656 lbs), make up 42% of the total. SWaN records no advance credits for nutrients. Observed

Demand

The nutrient market records no requirement, so demand cannot be measured directly (Section 02.5). Two things show where it arises. One is purchases: in the last four quarters, buyers bought 2,295 lbs of phosphorus. The other is construction: 7,360 construction stormwater (VAR10) permits were open on 30 September, and each may need credits under its local stormwater plan. Neither says how much is still needed. Observed

Where Supply Meets Purchases

Statewide, the phosphorus for sale equals about 2.6 years of purchases at the rate of the last four quarters. Locally, the trading rules matter. Construction sites may use credits generated in their own watershed or an adjacent one, so the supply within reach of a buyer includes the neighbouring watersheds. Counted that way, every watershed where phosphorus was bought has supply within reach, and in all but two it equals more than a year of purchases in the same area. The two are York and Nottoway, which together account for about 1% of purchases. Derived

That explains the busiest watershed. This quarter, 78 of the 89 purchases in Middle Potomac-Anacostia-Occoquan came from banks in other watersheds, mainly Lower Potomac (42) and Middle Potomac-Catoctin (28). Both border it, so these purchases are what the adjacency rule allows. Two conditions are not reflected in these figures: construction sites may use only perpetual credits, and tighter local conditions can apply where waters are impaired. Both could narrow the supply available to some buyers. Derived

How Supply and Demand Are Moving

Phosphorus credits for sale ended the quarter 2.8% lower than on 3 August, at 5,960 lbs, after peaking at 6,202 lbs in mid-August. Purchases were a little below recent levels: 551.8 lbs this quarter, against an average of 628 lbs over the previous four quarters. The register does not date nutrient bank approvals, and construction permits carry no registration date, so new supply and new demand cannot be traced quarter by quarter as they can for stream and wetland credits. Observed

Product Mix

Every one of the quarter's 410 phosphorus sales also included nitrogen, and 395 (96.3%) also included sediment. Since the start of 2024, every phosphorus sale has included nitrogen, and the share that also included sediment has risen from 77% to 96%. This is why the report counts sales by phosphorus alone: nitrogen and sediment come with it.

None of the quarter's 410 phosphorus sales has a recorded price. Prices for nutrient credits become known mainly when a public body buys, because Virginia's procurement system, eVA, publishes the purchase. SWaN matches those eVA records to sales. The prices added in recent months all relate to earlier sales; the most recent is from October 2025. Observed

Across the 81 phosphorus sales with a recorded price, from 2017 to October 2025, the middle price was $13,000 per pound, and half of all prices fell between $11,743 and $16,000. The 11 prices from 2024 and 2025 ranged from $9,599 to $20,541, with a middle price of $14,000. Almost all the buyers were public bodies: VDOT, universities, colleges and state agencies. These prices show what public buyers have paid. They may not reflect what private developers pay. Observed

Unlike the stream and wetland market, the nutrient market has no in-lieu fee schedule to act as a guide, so these public purchases are the only price evidence available.

Chapter 1080 (HB 952)

Industrial sites with stormwater permits must reduce the nutrients they release. Chapter 1080 lets them meet their 2025 and 2026 limits with "point source" credits, which are produced by regulated facilities such as wastewater plants. Conditions apply. The credits must be produced and used in the same year, and must come from the same locality or river. Credits for 2025 had to be bought by 1 September 2026, during this quarter. Credits for 2026 must be bought by 1 June 2027. The law also tells DEQ to focus enforcement on sites that do not comply, and to report on future arrangements by 1 November 2026. Observed

What the Record Shows

Point source credits are not sold through the nutrient banks covered here. So if industrial sites switched to them, the only sign would be fewer bank purchases by those sites. This quarter, 2 phosphorus sales named an industrial stormwater permit, compared with 2 to 4 a quarter since late 2025. That is too few to show any effect. Observed DEQ's report, due on 1 November, should say more.

A Market Without Advance Notice of Demand

The nutrient market gets no advance notice of demand, and that did not change this quarter. Each construction project's phosphorus requirement is set in its local stormwater plan, one of about 90, and is not published. DEQ's construction permit records, now in SWaN, do not include it. They record the phosphorus bought only when the permit ends, often long after the purchase. And 388 of the quarter's 410 phosphorus sales name no permit at all. So demand only becomes visible when a sale is made. Observed

This means buyers and sellers see different things. Nutrient banks put land and money into producing credits without knowing where or when they will be needed. Their decisions rest on past sales, not on known future need. Buyers show their demand only when they buy. The stream and wetland market works differently. Each VWP permit states its requirement when it is issued, so banks can see demand building in a watershed before anyone buys. Implied This difference comes from how data is collected to meet regulatory responsibilities in each market. It is not a fault in SWaN's records.

Several of the quarter's identifiable buyers bought in both markets: VDOT, Dominion Energy, Chesterfield County, Spotsylvania County, the City of Chesapeake, Lennar and Stanley Martin. Q0 found 168 organisations that bought in both markets between 2021 and 2025. Roads, utilities and housing often damage streams or wetlands and also add hard surfaces that increase runoff. So they need both kinds of credit at once, and one cannot replace the other.

This quarter's two regulatory changes affect the markets separately. The federal proposal affects only the federal side of the stream and wetland market. The new state law affects only industrial sites in the nutrient market. A project that needs both kinds of credit still needs both, but the rules behind them are now less alike. Implied

Neither market recorded a sale price this quarter. The stream and wetland market at least has a guide in the in-lieu fee programme's fees: three were recorded, from $60,000 to $100,000 per acre of wetland. The nutrient market has no such guide. In both markets, purchases by public bodies are the most likely source of real prices. This quarter's public buyers included Virginia Passenger Rail Authority, VDOT, counties and school boards.

The two markets differ most in whether demand is known before anyone buys. In the stream and wetland market, each VWP permit states what it needs. From this quarter, SWaN can read those needs from DEQ's record and compare them with credits for sale in each watershed. It can also follow them over time. So far, though, the figures are not yet complete. The nutrient market has nothing like this. Its demand shows up only when sales are made, and this quarter even the sales named fewer buyers. Both markets show what is for sale. Only the stream and wetland market shows what is needed. Derived

North Carolina

SWaN also holds records for North Carolina's stream, wetland and nutrient markets. This report does not analyse them, but they show how far Virginia's sellers operate across the state line. Six organisations own or act for banks in both states. Resource Environmental Solutions is the largest in both, with 117 banks in Virginia and 82 in North Carolina. Wildlands Engineering (80 banks) and Restoration Systems (45) are mainly North Carolina businesses with a small presence in Virginia. Over the last four quarters, four of the six sold stream or wetland credits in both states, and RES made the same number of sales in each: 41. Observed

Three limits apply. North Carolina's records hold no permit requirements, so there is no demand side to compare. Its credit types differ from Virginia's: streams are split into warm-, cool- and cold-water credits, wetlands into riparian and non-riparian, and nutrient credits include buffer credits. So amounts cannot be added to or compared with Virginia's. And SWaN's record of North Carolina nutrient trades falls from 91 in January–March 2026 to 11 in April–June and none in July–September; the data does not show whether that reflects trading or how North Carolina's records reach SWaN. In the quarter, SWaN recorded 28 stream and wetland trades in North Carolina, against 75 in Virginia. Observed

The table below brings together the main measures in this edition. From the next edition, MNM will report them every quarter, so that changes in the market can be followed over time.

A SWaN Index is in development. It will read market conditions where prices cannot be seen: whether credits are available where they are needed, whether there is a backstop where banks fall short, how concentrated local markets are, whether demand can be seen coming, and whether prices are visible. It will be introduced in the next edition, once its rules have been settled and tested against a second quarter of data.

Data

This edition uses a full copy of the SWaN database taken on 1 October 2026. SWaN last updated from RIBITS and from DEQ's permit records on 30 September 2026.

It covers 36,996 live trade records for Virginia. Credit amounts are stored multiplied by 1,000, and are converted back. "Commercial" means banks that sell to any buyer. Single-client banks and in-lieu fee purchases are reported separately. A nutrient sale is counted once, by its phosphorus credit.

We leave out 2,644 phosphorus purchases that were loaded from an earlier DEQ export of construction stormwater credits. They have no selling bank, may repeat sales already in RIBITS, and are under review. None of them falls in this quarter or the four quarters before it.

Credits for sale and credits owed come from SWaN's daily records (3 August to 1 October 2026). Bank details come from the bank register. Sellers are shown by the account that registered the bank. Banks that cannot yet be linked to an owner are counted as separate sellers in the nutrient market. Buyers are shown by organisation where SWaN can identify one.

Seller concentration is measured with the Herfindahl–Hirschman Index (HHI): the sum of each seller's squared percentage share of the quarter's trades, or of the amount sold where stated. Sellers are counted by registered account, and also by bank for comparison.

Local Markets and the Trading Rules

Local measures follow the trading rules rather than treating each watershed as a market. For stream and wetland credits, a watershed counts as served by a bank when at least half of it lies within the bank's primary or secondary service area. SWaN holds 593 service-area records (548 primary and 45 secondary). They cover 86 of the 104 approved commercial banks; the other 18 hold only 125 LF of stream credits. With a looser rule, counting a watershed as served when a tenth of it lies within a service area, the stream watersheds still short fall from four to three, holding 18% of credits owed. Non-tidal wetland credits by bank come from the credits recorded for each bank, with tidal banks excluded. These total 578 acres, against 416 acres in the snapshot, so wetland figures by service area should be read as indicative. In-lieu fee advance credits are recorded for each programme as a whole and are not placed by service area.

For phosphorus, the supply within reach of a watershed is the supply in that watershed and its neighbours, following Code of Virginia §62.1-44.15:35. Two further conditions are not applied: construction sites may use only perpetual credits, and DEQ can require credits from closer to the site where local waters are impaired (9VAC25-900-91).

Reporting Delays

Trades reach SWaN through weekly updates. For trades in 2026 that arrived after 22 July, the typical delay was 6 days for nutrient sales and about 11 days for stream and wetland sales. Between 27% and 43% arrived more than two weeks late. In-lieu fee advance sales arrive in batches, often many months after the sale: of the 26 recorded since late July, 20 arrived together on 3 August, covering sales from April 2025 to March 2026.

DEQ Permit Records

On 30 September SWaN replaced DEQ's periodic exports with records drawn through DEQ's API. Table N-1 shows what was loaded.

The load is marked partial because not every permit can yet be matched to trades (see below). Two of the 10,576 permits DEQ published also could not be saved, WP3-24-0171 and WP4-25-0831, because their contact details failed a format check. Separately, 27 records were saved without a watershed, because DEQ published neither a watershed nor usable map coordinates for them. They do not appear in SWaN's permit lists. Observed

Matching DEQ's record to purchases

In August 2026 Water Ledger checked DEQ's stream and wetland permit data (4,487 rows) against RIBITS, permit by permit. First, each permit was graded on how far its own record can be trusted. Then the most reliable permits were matched to purchases in RIBITS. The match was tried first by permit number, then by exact buyer name, then by a looser name match in the same watershed within 90 days. Table N-2 shows the results, in credits as DEQ records them.

Three things follow.

First, only the confirmed part of the record can be stated as fact. That means permits with clean records whose purchases match by permit number. On that part, 50,175 stream credits and 133.7 wetland credits are still owed.

Second, the rest of the record is unconfirmed, not unpaid. A permit with no match may have been paid for under a different number or name. Records in acres or feet cannot be turned into credits without a conversion rate. And DEQ's own "Open" label is not reliable: 47 stream permits marked Open already have a matching purchase. For 31 of them the purchase is within one credit of the requirement, and it was made a median of 2.3 years earlier.

Third, taken at face value, the whole record shows about four times the confirmed stream demand and seventeen times the confirmed wetland demand. How much of that is real, still owed and convertible is not yet known.

SWaN handles stream and wetland permits in two steps. First, it loads a permit, with its full requirement, only if it has full confidence in the permit's details from DEQ's API. In the August review, 1,561 of the 1,770 stream and wetland requirements DEQ recorded (88.2%) met this standard: 579 of 702 for stream (82.5%) and 982 of 1,068 for wetland (91.9%). Of the rest, 195 were recorded in units that cannot be converted to credits without a rate (116 stream, 79 wetland), and 14 had other problems, such as inconsistent figures or no quantity. Permits that fail the check are not loaded yet. Their demand is missing, so total demand is understated. They are being checked with DEQ. Once a permit is corrected and passes the check, it is loaded in SWaN's weekly automatic update.

Construction stormwater (VAR10) permits

DEQ records no credit requirement for VAR10 permits. It records only the pounds bought, the selling bank and whether a sale affidavit was received, and only when a permit ends. So for these permits, confidence is about the recorded purchases, not about demand. In September 2026 Water Ledger checked each purchase DEQ recorded against the selling bank's own ledger in RIBITS. Table N-3 shows the results.

Of the 812 permits that record a purchase, 692 (85%) are confirmed by the bank's ledger. For 106, no matching sale can be found; for 2 of these the difference may be rounding. Of the 14 with an unidentified bank, 9 name a known bank under a different spelling. The other 5 cannot be placed: four name a company rather than a bank, and one names no bank on record. The 7,272 permits with nothing stated are mostly not a data problem. DEQ records a purchase only when a permit ends, so a permit still running correctly shows nothing.

The API has other gaps. DEQ publishes 9,150 rows for 8,084 permits, because 839 permit numbers have two to seven rows. These are the permit's approval history, not extra purchases: no permit states two different amounts or two banks. The effective-date fields hold the same fixed value (1 July 2024) on every row, so the approval-signed date is used to find the latest row. The listed watersheds disagree with each other for 353 permits; SWaN uses the published map coordinates instead. Of the remaining gaps, 327 permits list more than one operator, 18 have neither a watershed nor coordinates and cannot be processed, 319 have no date of any kind, and 28 have no operator email. The pull also contains no statement of which nutrient the pounds refer to. All are treated as phosphorus, which DEQ has been asked to confirm. Observed

The 30 September load contained 8,211 VAR10 permits, slightly more than the September review. Of the 2,644 VAR10 permits in DEQ's earlier export, 486 still appear in the API, with identical amounts. The other 2,158 no longer appear. The API also shows 326 permits with purchases that the earlier export did not cover.

Phosphorus purchases linked to permits

When the earlier records were retired, the phosphorus purchases SWaN links to permits fell from 6,560.8 lbs to 25.6 lbs. No purchases were lost. SWaN counts a purchase against a permit only when the permit has a credit requirement. The earlier VAR10 permit records each had one, taken from DEQ's export, and the 6,560.8 lbs was counted against them. DEQ's own VAR10 records have no requirement. They record only the pounds bought, and only when the permit ends. So when the earlier records were retired, their purchases had no requirement to be counted against.

The purchases themselves are all still in SWaN. The 25.6 lbs left belonged to 11 earlier permits that were kept because their owners had confirmed them. This result was expected before the change was made. Observed

Corrections to Q0

Changes of Basis

Linking Sales to Permits

We measure what is owed from DEQ's record, not by counting sales. A sale is linked to a permit only when the permit number on the sale matches DEQ's record. Across all years, 8.8% of stream and wetland sales match exactly. For this quarter's sales the figure is 16.7%, so recent sales are easier to link. The platform's more detailed matching method raises the all-years figure to 13.2%.