Research brief · Applied Energy

A gas field, already plumbed.

Every U.S. wastewater plant with a digester makes methane. Feed it hydrogen and it makes pipeline gas. We modelled all 2,952 of them to find where that pays.

Jiang J., Du L., Li X., Chen C., Ren Z.J. Spatially resolved national assessment of renewable natural gas potential from water resource recovery facilities through biomethanation. Applied Energy 417 (2026) 128025. Read the paper →

stays as-is new pipeline trucked gas
Every facility with anaerobic digestion in the contiguous U.S., placed on an equal-area projection. Point size follows treatment capacity; the interactive version is further down.

What the model found

A national database of every wastewater facility running anaerobic digestion — assembled from CWNS, WEF, DOE CHP and Argonne National Laboratory records — put through a facility-by-facility optimisation.

Facilities modelled
2,952
Every plant with anaerobic digestion in the contiguous U.S.
Economically viable
46.4%
1,369 plants clear the bar under base capital cost.
Renewable gas capacity
5,781 MW
Upgrading biogas to pipeline-grade methane.
Net annual benefit
$2.2B
Summed across all viable facilities, per year.

One optimisation, run 2,952 times

Biomethanation converts digester CO2 into methane using hydrogen. That hydrogen has an alternative buyer, the upgraded gas has to reach a pipeline, and both facts vary by location — so the decision is different at every plant.

A mixed-integer program was written for each facility, carrying its own energy prices, its own distance to gas infrastructure, its treatment capacity and the incentives available in its state. Each was solved to the pathway with the best annualised cash flow against business-as-usual, with capital cost piecewise-linearised and an industry-guided 70% producer share assumed on RIN revenue.

  1. Build the database

    CWNS, WEF, DOE CHP and Argonne records reconciled into 2,952 facilities.

  2. Locate every plant

    Pipeline distances, state energy prices and LCFS eligibility attached to each site.

  3. Solve for the best pathway

    Mixed-integer program per facility, solved globally in Gurobi.

  4. Shake the assumptions

    1,000 Monte Carlo iterations across seven uncertain parameters.


Four pathways, one winner per plant

Each facility is assigned the pathway that maximises its annualised cash flow relative to doing nothing. Under base capital cost, the CHP upgrade never wins anywhere.

PathwayWhat happensPlants
Stay as-is

Business as usual. No biomethanation; biogas burns in the existing combined heat and power unit, and any electrolysis runs independently.

1,583
Upgrade the CHP

Biomethanation built, upgraded gas fed back into the CHP for more heat and power. Only viable under low capital cost — 119 plants, 91% of them in California — and never chosen at base or high cost.

0
Build a pipeline

Renewable gas injected into the grid through newly built pipe to the nearest gas infrastructure.

432
Truck it instead

Gas compressed and hauled by CNG trailer to an injection point — the answer for remote and mid-size plants.

937

Every plant, and what it should do

All 2,952 facilities under base capital cost. Filter to see where each pathway wins.

Showing 2,952 facilities. Circle size follows treatment capacity.

Drag to pan, scroll or use +/− to zoom. Boundaries: U.S. Census / Natural Earth (public domain).

Six things the run showed

01

Size decides who plays

1,369 of 2,952 facilities are viable at base capital cost, carrying 5,781 MW of renewable gas capacity. The cutoff sits at 2–3 MGD; above 10 MGD, almost every plant clears it.

02

RIN credits decide everything else

Renewable identification numbers bring $4.69B a year across viable facilities — the largest single cash flow in the model, and the strongest positive driver of viability at PRCC +0.91.

03

Hydrogen is the thing you give up

Diverting hydrogen away from direct sale costs $1.31B a year. At PRCC −0.96, the hydrogen price is the most influential parameter in the whole analysis.

04

Trucks beat pipe, two to one

937 facilities — 68% of the viable set — choose CNG trucking over building pipe. New pipeline only pays when the plant is large enough or the run short enough.

05

State credits help, but don't carry

Low carbon fuel standards in California, Oregon and Washington add $145M a year. Real money, but not enough on its own — federal incentives stay essential.

06

The result holds when shaken

Across 1,000 stochastic runs, viability lands between 21.5% and 59.6%, median 44.2%, with a 90% interval of 34.5–52.2%. The conclusion is structural, not a knife-edge.


What moves viability

Partial rank correlation coefficients from 1,000 Monte Carlo runs. Bars right raise the share of viable plants; bars left lower it.

Revenue on one side, hydrogen on the other

Each facility's annual economics balance renewable gas sales, RIN credits and LCFS credits against capital cost, pipeline construction, forgone CHP output, and the hydrogen it no longer sells.

The two parameters that matter sit at opposite ends of the chart, and both are prices set outside the plant gate. That is the finding underneath the finding: this is not an engineering problem waiting on a better reactor, it is a market problem waiting on a stable signal.

Where the plants are

Facilities by pathway for the top 30 states under base capital cost.


Shaking seven assumptions at once

A thousand runs, each drawing new values for hydrogen price, RIN price, natural gas price, electricity price, discount rate, trucking cost and pipeline capital cost.

How often it works

Share of facilities viable in each run. Median 44.2%, 90% interval 34.5–52.2%.

And what breaks it

One dot per run. Colour carries the RIN price coefficient — blue high, rust low.


Three levers, and only one of them is load-bearing

Federal and state incentives reshape this landscape more than any technical parameter. The analysis assumes producers keep 70% of RIN revenue.

Federal RIN credits

PRCC +0.91

  • D3 credits generate $4.69B a year across viable facilities.
  • The strongest positive driver of viability in the model.
  • Strip them out and almost no facility reaches positive cash flow.
  • Price swings dominate: at the low end under 10% of plants are viable, at the high end over 60%.

State LCFS programmes

$145M per year

  • Active in California, Oregon and Washington.
  • Supplementary revenue for eligible facilities — meaningful, not sufficient.
  • Extending the programmes to more states would widen the viable pool.

The hydrogen trade-off

PRCC −0.96

  • Forgone hydrogen sales cost $1.31B a year — the largest cost in the model.
  • The single most influential parameter overall.
  • Base case assumes hydrogen sells at $2/kg; as green hydrogen gets cheaper, biomethanation gets easier.

Policy is the kingmaker

Biomethanation at wastewater plants is a scalable, infrastructure-compatible route to renewable natural gas, and the economics are not subtle: RIN credits decide whether projects happen, and hydrogen opportunity cost decides where they stop. Across 1,000 simulations viability holds at a median of 44.2%, with a 90% interval of 34.5–52.2%. What municipal investors need is not a better process — it is a RIN market that stays put long enough to finance against, or capital grants that do the same job.

Read the full study →