United States: for consulting engineers

Where the work is, what is being specified, and what the regulatory calendar actually says.

What is driving spend

We read 5,677 Clean Water SRF project entries across 39 states and matched them to live NPDES permits for design flow. In practice these intended use plans and priority lists name the applicant, the project and the amount, and most carry some indication of readiness. That makes them the closest thing there is to a public schedule of who is about to build.

Overlaying that with EPA's quarterly non-compliance record gives the set that actually matters: 81 plants that are lagoon-based, inside 5 MGD, queued for funding, and carrying reported effluent limit violations in the last three years.

For a plant in this band the municipality rarely selects the process. You do. So the useful intelligence is not which town is failing. It is where the work is concentrated and what the alternatives analysis has to survive.

The project descriptions are the intelligence

SRF priority lists carry a narrative description per project, and it is far more useful than the columns. One state's top-ranked project states plainly that the plant "utilizes a lagoon treatment system" and that the upgrade replaces it with biological nutrient removal followed by a membrane bioreactor.

That single sentence gives the existing plant type, the selected process, and therefore what was displaced. Read across 5,677 projects it becomes a picture of what is actually being specified at this size, as opposed to what the literature says should be.

What is in the addressable band

2,102 of the projects we matched sit inside 5 MGD design flow, or 1,595 distinct plants, since a town financed in two cycles appears twice. That is the band where a conventional nutrient upgrade is hardest to justify per capita, and where the alternatives analysis is most often genuinely open.

Where the load is coming from before it reaches the works

We measured land cover within 10 km of every matched plant from USDA's Cropland Data Layer. 649 sit in catchments at least 40% corn and soybeans, the nitrogen-fertilised rotation the Gulf hypoxia and Lake Erie programmes are written around. Combine that with the receiving-water class and 122 plants have all three: a small receiving water, a catchment in drought, and violations already on file.

Two cautions worth carrying into an alternatives analysis. First, EPA's CAFO permit file is a floor and not a census: only about 6,760 operations hold an NPDES CAFO permit against roughly 21,000 large operations nationally, so a zero there means unpermitted, not absent. Second, agricultural context is a reason the receiving water is already close to its limit. It is not an argument that any particular works is the cause, and treating it as one is how these conversations get lost.