The same plant, read twice.

Nobody chose this

A mine opened in northern New Mexico about a century ago because the ore was wanted, and it ran, on and off, for generations. It was the village's employer. It is now a Superfund site under a court-approved cleanup, and the company that closed it is still there, treating water and covering tailings under an obligation that outlives the ore.

Then a new industry arrived with a promise that fits the wound almost too neatly: build clean energy on the ground extraction ruined, and give the place its employer back. That promise is not cynical. The federal tax code pays a premium for it. A member-owned cooperative — the same one that bought its way out of its wholesale power contract so it could choose its own supply — is the one carrying it forward.

And the hard parts are structural, not personal. A basin is over-appropriated before anyone applies. A federal review can clear a project without ever quantifying the one thing the village is asking about. A standard can ask for consultation and call it enough. A sector can publish gigawatts and deliver a rounding error. None of that requires a villain. It only requires that each step be reasonable on its own terms while nobody is accountable for the whole.

Each line below is an outcome first — something a place would actually notice — and underneath it the instrument that produces it. The badge says how proven that instrument is, and the badge never comes off.

Operating means it is running somewhere and measured. Pilot means it works at small scale. Standard means it is established practice. Regulatory means the law already requires it in some jurisdiction. Proposal means exactly that: this model is asking for it, and cannot point to it being done.

Thirteen of the fifteen badges below are the tier carried by that instrument's own row in this model's pattern library, and the slug names the row, so a reader can check any badge against the sources that produced it. The last two name a rule this model imposes on itself rather than a library instrument; their figures are marked as modelled rather than sourced, and they are counted as unproven in the note at the foot of the list.

A reference model cannot tell a valley how much water it can spare. It can make sure that the people who live on that water have the measurements, the standing and the money to answer it themselves — and that the answer, whichever way it goes, is binding.

So the sequence is not: choose the technology, then consult. It is: read the place, ask the people who can refuse, measure the water, name what the hydrogen is for, and only then decide whether there is a project here at all. A no reached that way is a good outcome. It is the same standard working.